- Email: [email protected]
Plasma and urinary free amino acid concentrations in preeclamptic women in northern Nigeria
Clinica Chimica Acta 342 (2004) 179 – 185 www.elsevier.com/locate/clinchim
Plasma and urinary free amino acid concentrations in preeclamptic women in northern Nigeria Robert H. Glew a, George Melah b, Aliyu I. El-Nafaty b, Yekaterina Brandt c, D. Morris c, Dorothy J. VanderJagt a,* a
Department of Biochemistry and Molecular Biology, MSC08 4670, 1 University of New Mexico School of Medicine, Albuquerque, NM 87131-0001, USA b Department of Obstetrics and Gynaecology, Federal Medical Centre, Gombe, Gombe, Nigeria c Department of Surgery, University of New Mexico School of Medicine, Albuquerque, NM 87131-0001, USA Received 18 November 2003; received in revised form 29 December 2003; accepted 31 December 2003
Abstract Background: Preeclampsia/eclampsia is one of the major causes of maternal and fetal mortality in Nigeria and many other countries in sub-Saharan Africa. We determined if serum and urine concentrations of amino acids were abnormal in women with this disorder of pregnancy in Gombe, Nigeria. Methods: Free amino acids were measured in serum and urine of women (ages, 15 to 40 years) with preeclampsia/eclampsia (n = 37) and their pregnant age and gestational age matched controls (n = 16). Results: The concentrations of 19 of the 20 serum amino acids that are common in proteins were not significantly different between the control and preeclamptic groups. Phenylalanine was increased in the preeclamptic women compared to the controls ( p < 0.05); however, the difference between the two values was small (119 vs. 104 umol/l, respectively). The concentrations of urine free amino acids were indexed to urine creatinine (i.e., nmol/mg creatinine). Five amino acids were significantly elevated in the urine of preeclamptic patients vs. the control group: proline (100%, p < 0.05), the branched amino acids, valine (80%, p < 0.05), leucine (61%, p < 0.05), isoleucine (49%, p = 0.01), and methionine (52%, p < 0.005). On the other hand, the concentrations of a nutritionally non-essential amino acid, glycine ( p = 0.01), and an essential amino acid, histidine ( p = 0.01), were both reduced by about one-third in the urine of the preeclamptic women. Conclusions: Our findings indicate that there are disturbances in urinary amino acid excretion but not in the serum amino acid profile in the preeclamptic patients. The degree of aminoaciduria is probably not sufficient to significantly affect overall nitrogen balance or precipitate a deficiency of one or more of the essential amino acids. D 2004 Elsevier B.V. All rights reserved. Keywords: Preeclampsia; Hypertension; Proteinuria; Aminoaciduria; Pregnancy; Nigeria
1. Introduction
* Corresponding author. Tel.: +1-505-272-5799; fax: +1-505272-3518. E-mail address: [email protected] (D.J. VanderJagt). 0009-8981/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.cccn.2003.12.028
Preeclampsia/eclampsia is a major cause of maternal and perinatal mortality in northern Nigeria [1,2] and many other regions of sub-Saharan Africa. It is the leading cause of maternal mortality at the Specialist Hospital in Maiduguri, Nigeria [3]. Although the risk of a pregnancy developing into eclampsia can be
180
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
reduced by good antenatal care, there is still no method of preventing preeclampsia. Furthermore, although there is a growing literature pointing to diffuse endothelial dysfunction in preeclampsia/eclampsia [4], the cause of this multisystem disorder of pregnancy remains elusive. Preeclampsia usually occurs who have been normotensive after the 20th week of pregnancy and is characterized by hypertension, with or without edema, and proteinuria. The disturbances associated with preeclampsia include: hyperlipidemia, particularly hypertriglyceridemia [5,6]; excessive lipid peroxidation and insulin resistance [7]; and imbalances in the placental production of thromboxane A2 and prostacyclin [8]. There is a lack of consensus in the literature regarding amino acid concentrations in the blood of preeclamptic women. Li et al. [9] reported finding increased concentrations of aromatic amino acids, threonine, arginine, glycine, cystine and glutamic acid in the plasma of preeclamptic women. On the other hand, Kremenski et al. [10] found preeclampsia to be associated with increases in the plasma concentrations of aspartic acid, glutamic acid, threonine, glycine and asparagines, but not the aromatic amino acids or cystine. Contradicting the findings, D’Aniello et al. [11] reported a five-fold lower mean plasma arginine concentration in preeclamptic women relative to controls matched for age and stage of gestation. Moreover, these investigators found that the concentrations of most of the other 19 common amino acids were significantly reduced, on average by about 50%, in the serum of preeclamptic subjects relative to the controls. The present study was conducted in order to determine if there were abnormalities in the concentrations of the serum amino acids of preeclamptic women in northern Nigeria. Our hypothesis was that the proteinuria-based protein wastage associated with preeclampsia, together with the marginal protein intakes of the general population in northern Nigeria where this study was conducted, would cause the preeclamptic women to have reduced serum concentrations of one or more of the essential amino acids. We also took advantage of the opportunity to compare the urinary excretion of free amino acids between preeclamptic women and healthy pregnant controls.
2. Materials and methods 2.1. Subjects Pregnant women were recruited into the study at the Federal Medical Centre, Gombe and the Specialist Hospital, Gombe, Nigeria. The diagnosis of preeclampsia was made in 37 pregnant women based on the criteria proposed by the American College of Obstetricians and Gynecologists which included: cerebral or visual disturbances, epigastric pain, pulmonary edema or cyanosis, a systolic blood pressure z 140 mm Hg or a diastolic blood pressure z 90 mm Hg and proteinuria. We set the cut-off for proteinuria at 190 mg total protein/g creatinine [12]. Protein was estimated using the method of Bradford [13] with albumin serving as the protein standard. Serum uric acid concentrations were not determined. Ten milliliters of venous blood was collected into clean, dry vacutainers and allowed to clot for 45 min at room temperature. After centrifugation at 5000 g for 8 min, the clear supernatant was aliquoted into 2-ml cryovials and stored for 2– 6 weeks at 43 jC until which time the specimens were transported in the frozen state to Albuquerque for analysis. At the same time blood was drawn, urine was collected into sterile plastic cups, aliquoted into cryovials and stored and transported as described above for blood serum. The control population consisted of 16 healthy pregnant women whose mean age and gestational age matched the corresponding mean values of the preeclamptic subjects. The study was approved by the Human Ethics Committee at the Federal Medical Centre, Gombe and the University of New Mexico School of Medicine, and informed consent was obtained for all subjects. 2.2. Amino acid analysis A 50-Al aliquot of serum or urine was mixed with an equal volume of 0.5 mmol/l norleucine, the internal standard. Norleucine was used as the internal standard because it is not commonly found in proteins. For the purpose of deproteination, the mixture was pipetted into an ultrafiltration device (Ultrafree-MC, Milllipore, Bedford, MA) and centrifuged at 2400 g for 30 min. No special precautions were taken with regard to tryptophan. An aliquot of the supernatant was transferred into a 6 50-mm glass tube and dried in a
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
vacuum. Following drying, 10 Al of redry solution (methanol/ 1.0 mol/l sodium acetate/ triethylamine, 2:2:1) was added after which the sample was redried. Next, 20 Al of phenylisothiocyanate (PITC) reagent (methanol/water/triethylamine/phenylisothiocyanate, 7:1:1:1) was added under alkaline conditions. The reaction was allowed to proceed for 20 min at 20 jC to produce phenylthiocarbamoyl amino acids (PTCAA) [14]. Excess reagent was removed at 20 jC with the aid of a vacuum pump. After drying the sample, the PTC-AA were dissolved in 0.1 ml buffer A (0.07 mol/ l sodium acetate, pH 6.5 containing 2.5% (v/v) acetonitrile and 1 ppm EDTA, Ca+ 2 salt). The derivatized amino acids were then separated by reverse-phase high performance liquid chromatography using an Ultra WISP 715 autoinjector and a Waters Pico Tag solvent delivery system (Model 510). Elution of PTC-AA was achieved using gradient consisting of buffer A and buffer B (acetonitrile/water/methanol, 45:40:15, by volume) over 64 min at a flow rate of 1 ml/min [15]. The column effluent was monitored at 254 nm and the signal was integrated using Millenium 1999 ver. 3.20 (Waters, Milford, MA) software. The retention times of the various amino acids were determined using an amino acid calibration mixture. Quantitation of individual amino acids was achieved by monitoring the absorption of the column eluate at 254 nm and comparing the areas under the individual peaks with those of the corresponding amino acid standards. The average standard error of the mean for the 20 amino acids was 3.1%.
181
Table 1 Comparison of characteristics of preeclamptic/eclamptic women and their matched healthy controls from Gombe, Nigeria Parameter
Matched controls, mean F S.D. (n = 16)
Age (years) BP, systolic (mm Hg) BP, diastolic (mm Hg) Parity Gestational age (weeks) Urine protein/ creatinine ratio (mg/g)
26.9 (7.2) 107 (7) 69 (8) 2.5 (2.8) 35.3 (5.5) 121 (44)
Preeclampsia/ eclampsia, mean F S.D. (n = 37)
p-value
25.0 (7.3) 151 (25)
NS < 0.001
103 (14)
< 0.001
2.3 (2.9) 35.4 (3.9) 1115 (2308)
NS NS < 0.001
NS, non-significant ( p>0.05).
significant differences in the two groups in terms of parity (2.3 –2.5) or gestational age (approx. 35 weeks). The mean diastolic and systolic blood pressures of the preeclamptic women and their total urine total protein excretion (expressed as mg protein/g creatinine) values were significantly increased relative to the pregnant controls, which is expected given that blood pressure and urinary protein concentrations were the factors distinguishing the two populations. The mean gestational age of both control and preeclamptic subjects was approximately 35 weeks. The controls were selected to match the gestational age of the preeclamptic patients.
2.3. Statistical methods
3.2. Serum amino acids
The two-sample t-test was used to compare variables between patients with preeclampsia and controls using the NCSS 2001 Statistical System for Windows (Kaysville, UT). A p-value < 0.05 was considered significant.
Table 2 shows that, with the exception of phenylalanine, the serum concentrations of the common amino acids were not significantly different between the pregnant controls and the subjects with preeclampsia. Furthermore, even though the phenylalanine concentration was significantly higher in the preeclamptic women compared to the controls, the difference between the two values was < 20%. The serum total amino acid concentration was not significantly different between the control and preeclamptic populations (3030 vs. 3150 Amol/l). Noteworthy was our finding that, in contrast to the study by D’Aniello et al. [11] which reported a five-fold reduction in the plasma arginine concentration of preeclamptic women, we
3. Results 3.1. Comments on the subjects As shown in Table 1, the mean ages of the preeclamptic women and the pregnant controls (25.0 and 26.9 y) were not significantly different, nor were there
182
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
Table 2 Comparison of the concentrations of amino acids in the serum of preeclamptic/eclamptic women and matched controls (Amol/l) Amino acid
Matched controls (n = 16), mean F S.D.
Preeclampsia/eclampsia (n = 37), mean F S.D.
Asp Glu Ser Asn Gly Gln His Thr Ala Arg Pro Tyr Val Met Cys Ile Leu Phe Trp Lys Total
62 (18) 236 (62) 214 (46) 114 (22) 278 (62) 438 (101) 47 (25) 92 (32) 438 (110) 148 (110) 217 (64) 57 (12) 173 (43) 24 (5) 8 (3) 55 (15) 175 (34) 104 (23) 7 (6) 141 (54) 3030
63 (15) 242 (101) 218 (40) 107 (28) 308 (78) 455 (115) 48 (23) 107 (44) 451 (121) 151 (87) 221 (62) 59 (16) 180 (36) 25 (7) 8 (4) 54 (16) 175 (41) 119 (25)a 8 (8) 147 (46) 3150
a
p < 0.05.
found the serum arginine concentrations of the pregnant controls and preeclamptic women to be essentially the same. 3.3. Urine amino acids The urine concentrations of amino acids were expressed on the basis of urine creatinine (i.e., nmol/mg creatinine). In contrast to serum where the free amino acid profiles there were nearly indistinguishable, a number of relatively large and statistically significant differences were found when we compared the concentrations of the 20 common amino acids in the urine of the preeclamptic and control pregnant women. The concentrations of two nutritionally non-essential amino acids, namely glycine and histidine, were both reduced by about one-third in the urine of the preeclamptic women. On the other hand, the following five amino acids were present in increased amounts in the urine of the preeclamptic subjects relative to the controls: proline (100%, p < 0.05); the branched amino acids, valine (80%,
p < 0.0.05), leucine (61%, p < 0.0.05) and isoleucine (49%, p = 0.01); and methionine (52%, p < 0.005). No correlations were found when the urine concentration (indexed to creatinine) of proline, valine, leucine, isoleucine or methionine was plotted against the total urinary protein, also indexed to creatinine.
4. Discussion The present study yielded two main findings. First, the serum amino acid concentration profiles of preeclamptic women in Gombe, Nigeria were not very different from those of healthy pregnant women matched for age, gestational age and parity. Second, preeclampsia was associated with abnormalities in amino acid excretion. The type of aminoaciduria we documented in the preeclamptic women in our study was most likely of the secondary kind, as opposed to the overflow aminoaciduria one sees in fulminant hepatic failure [16]. It is noteworthy that four of the five amino acids that were present in increased amounts in the urine of the preeclamptic women are regarded collectively as belonging to the hydrophobic amino acid class. Glycine and the essential amino acid histidine were excreted in significantly reduced quantities relative to the controls. How might one account for the altered excretion of certain amino acids in preeclampsia? The excretion of greater than normal quantities of amino acids in the urine could be accounted for by a compromise of the function of one or more of the amino acid transporters in the proximal convoluted tubule of the kidney. There are five stereospecific reabsorptive systems for amino acids [17]: Group I, neutral monoamino, monocarboxylic and cyclic amino acids; Group II, cystine and dibasic amino acids; Group III, imino acids and glycine; Group IV, dicarboxylic amino acids; and Group V, h-amino acids. Most amino acids are reabsorbed by more than one transporter. Since methionine excretion was increased and histidine excretion was decreased in the urine of the preeclamptic subjects, and because both of these amino acids fall into the Group I class of transporters, a defect in one of the amino acid transporters in the kidney in the preeclamptic women in our study cannot account for the selective aminoaciduria we observed in women with this hypertensive disorder of pregnancy. By similar reasoning, since glycine and
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
proline are both Group III amino acids, our finding of decreased glycine excretion and increased proline excretion in the urine of the preeclamptic women relative to the healthy pregnant women argues against a hypothesis asserting that defects in one or more of the amino acid transporters in the kidney could account for the abnormal quantities of particular amino acids we observed in the urine of the preeclamptic women in Gombe. We have no alternative explanation for our amino acid findings. The results of the present study of urine concentrations of amino acids in preeclamptic women in northern Nigeria are not in accord with the results of a similar study by Velco et al. [18]: whereas we found abnormalities in the urinary excretion of seven amino acids in 37 preeclamptic women, they found no differences in the urine concentrations (indexed to creatinine) of amino acids in four women with proteinuric hypertension when compared with that in normal pregnancy. Our results regarding serum amino acid concentrations in preeclampsia also differ markedly from D’Aniello and coworkers’ [11], who reported a five-fold reduction in the mean plasma arginine concentration of 12 preeclamptic women relative to an equal number of pregnant controls matched for age and gestational age. The mean serum arginine concentrations of our preeclamptic and control pregnant were essentially the same (Table 2). Since hemodilution and an increase in the glomerular filtration rate are associated with pregnancy and both are reduced during preeclampsia, we expected to find significant increases in the serum concentrations of many amino acids in the preeclamptic subjects. A reduction in hemodilution in the preeclamptic subjects would tend to elevate amino acid concentrations, as would a decrease in the glomerular filtration rate. However, because we did not determine the concentration of albumin in the sera of any of the subjects in our study, we cannot comment on the degree to which hemodilution might have been a factor in affecting serum amino acid values of the preeclamptic subjects. Similarly, because we did not estimate the glomerular filtration rate of the subjects in our study, we are not in a position to comment on how this particular aspect of kidney function might have influenced our findings with respect to serum and urine amino acid concentrations.
183
The fact that we did not find noteworthy or significant differences in the concentrations of amino acids in the serum of preeclamptic vs. healthy pregnant women leads us to conclude that abnormalities in maternal amino acid metabolism were not associated with preeclampsia in northern Nigeria. Furthermore, this observation also suggests that transplacental transfer of amino acids from the mother to the fetus was probably not affected in the preeclamptic Nigerian subjects in our study. In contrast to our results, Evans et al. [19] recently reported finding increased concentrations of a number of amino acids in the serum of preeclamptic subjects in the US which they attributed to enhanced placental amino acid transport or reduced fetal amino acid utilization. The discrepancy between their findings and ours with respect to serum amino acid concentrations in preeclampsia may be accounted for by the fact that the preeclamptic subjects in our study were probably significantly less well nourished than the hypertensive, proteinuric subjects in their study. However, our data do point to disturbances in urinary amino acid excretion in the preeclamptic women we studied. We can only speculate regarding the mechanism that might account for the abnormal Table 3 Comparison of the concentrations of essential amino acids in the urine of preeclamptic/eclamptic women and controls (nmol/mg Cr) Amino acid
Matched controls (n = 16), mean F S.D.
Preeclampsia/eclampsia (n = 37), mean F S.D.
p-value
Asp Glu Ser Asn Gly Gln His Thr Ala Arg Pro Tyr Val Met Cys Ile Leu Phe Trp Lys Total
51.4 (38.7) 248 (72.4) 598 (189) 358 (169) 3450 (1511) 774 (299) 1090 (399) 210 (73.2) 651 (254) 17.4 (9.0) 23.1 (9.2) 103 (41.6) 47.8 (12.8) 22.0 (5.7) 28.0 (11.5) 27.0 (10.3) 65.1 (26.4) 73.7 (26.8) 78.9 (43.2) 139 (253) 8055
43.2 (43.3) 286 (187) 540 (341) 333 (267) 2382 (1654) 597 (373) 760 (579) 194 (189) 584 (377) 26 (25.6) 45.7 (55.2) 94.3 (59.1) 85.3 (77.2) 31.3 (16.0) 39.1 (28.4) 39.7 (18.4) 105 (73.0) 94.6 (53.9) 89.4 (70.6) 81.5 (78.9) 6451
NS NS NS NS < 0.05 NS 0.01 NS NS NS < 0.05 NS < 0.05 0.01 NS 0.01 < 0.05 NS NS NS
184
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
pattern of amino acids in the urine of these preeclamptic women. It is conceivable that the excessive protein in the urine of the preeclamptic women causes the uptake by proximal tubule cells of quantities of protein so large as to disturb the amino acid transport systems of these cells that are responsible for reabsorbing filtered free amino acids. However, since we did not carry out a thorough analysis of the nature or size of the proteins that were present in the urine of the preeclamptic women in our study, we cannot make a strong statement regarding the basis of the proteinuria they were experiencing. Much of the excessive urinary protein in the preeclamptic subjects in the present study can be explained by increased glomerular permeability. This deduction is based on the fact that in the 37 preeclamptic women, albumin accounted for 30– 70% of the total protein. It is well known that in glomerular proteinuria, the most abundant protein in urine is albumin [16]. However, since albumin did not account for all of the urinary protein in the preeclamptic women in our study, we suggest that defective tubular reabsorption also contributed significantly to the proteinuria they exhibited. In tubular proteinuria, albumin excretion is only slightly increased and proteins other than and of lower molecular weight than albumin usually account for most of the protein found in the urine [16]. We suggest that the mixture of albumin and non-albumin proteins in the urine of the preeclamptic women in the present study points to the co-presence of both glomerular and tubular proteinuria in these women. Finally, there is the question of what the detrimental consequence(s) might be of the increased excretion of certain amino acids by the preeclamptic subjects in the present study. Although the increased quantities of proline, valine, leucine, isoleucine and methionine excreted in the urine of the preeclamptic women were considerable, they would probably not be sufficient to significantly affect overall nitrogen balance or precipitate a deficiency of one or more of the essential amino acids. Since insulin plays such a key role in regulating the metabolism of the branched-chain amino acids, the fact that the quantities of valine, leucine and isoleucine were each increased by >50% in the urine of the preeclamptic subjects (Table 3) raises questions regarding insulin status and function in the preeclamptic women in our study.
In conclusion, although we have documented significant differences in the quantities of seven different amino acids in the urine of preeclamptic women, in light of the fact that the mean serum amino acid profile for this group of hypertensive, proteinuric subjects was, overall, so similar to that of the pregnant controls and because the degree of aminoaciduria in the patient group was only moderate, it is unlikely that differences in the urinary excretion of amino acids are of major clinical significance in the preeclamptic women we studied. Acknowledgements This study was supported by a Minority International Research Training (MIRT) grant from the Fogarty International Center of the National Institutes of Health. References [1] Harrison KA. Child-bearing, health and social priorities: a survey of 22 774 consecutive hospital births in Zaria, Northern Nigeria. Br J Obstet Gynaecol 1985;S5:1 – 119. [2] El-Nafaty AU. Pre-clampsia and eclampsia: a clinical guide for early detection and management for health workers in developing countries. Maiduguri: Ed-Linform Services; 1998. [3] Bobzom DN, Muna M. Maternal mortality at the State Specialist Hospital, Maiduguri, Borno State. Annals of Borno 1994;11/12:270 – 4. [4] Roberts JM, Redman CW. Preeclampsia: more than pregnancy-induced hypertension. Lancet 1993;341:1447 – 51. [5] Kaaja R, Tikkanen MJ, Viinikka L, Ylikorkala O. Serum lipoproteins, insulin and urinary prostanoid metabolites in normal and hypertensive pregnant women. Obstet Gynecol 1995; 85:353 – 6. [6] Gratacos E, Casals E, Sanllehy C, Cararach V, Alanos PL, Fortuny A. Variation in lipid levels during pregnancy in women with different types of hypertension. Acta Obstet Gynecol Scand 1996;75:896 – 901. [7] Walsh SW. Lipid peroxidation in pregnancy. Hypertens Pregnancy 1994;13:1 – 32. [8] Walsh SW, Behr MJ, Allen NH. Placental prostacyclin production in normal and toxemic pregnancies. Am J Obstet Gynecol 1985;151:110 – 5. [9] Li J, Sun J, Lu H. Determination of serum free amino acids and ammonia in pregnant women with severe pregnancy induced hypertension. Zhonghua Fu Chan Ke Zhi 1996;31: 468 – 70. [10] Kremenski I, Borison I, Barov D, Katsulov A. The plasma amino acid profile of women with a normal pregnancy and in pre-eclampsia. Akush Ginekol (Sofia) 1990;29:5 – 9.
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185 [11] D’Aniello G, Tolino A, Fisher G. Plasma L-arginine is markedly reduced in pregnant women affected by preeclampsia. J Chromatogr 2001;753:427 – 31. [12] Barratt M. Proteinuria. Br Med J 1983;287:1489 – 90. [13] Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein using the principle of protein dye-binding. Anal Biochem 1976;72:248 – 54. [14] Cohen SA, Strydom DJ. Amino acid analysis utilizing phenylisothiocyanate derivatives. Anal Biochem 1988;174:1 – 16. [15] Buzzigoli G, Lanzone L, Ciorciaro D, Fascerra S, Cerri M, Scandroglio A, et al. Characterization of a reversed-phase high performance liquid chromatographic system for determination of blood amino acids. J Chromatogr 1990;507: 85 – 93.
185
[16] Johnson AM, Rohlfs EM, Silverman L. Proteins. In: Burtis CA, Ashwood ER, editors. Tietz textbook of clinical chemistry. Philadelphia: W.B. Saunders; 1999. p. 477 – 540. Chapter 20. [17] Bonnardeaux A, Bichet DG. Inherited disorders of the renal proximal tubule. sixth edition. Brenner and Rector’s the kidney, vol. 2. Philadelphia: W.B. Saunders; 2000. p. 1656 – 98. [18] Velco CJ, Reynolds JW, Hawkins DF, Meek JH. Renal clearance of amino acids in pregnancy complicated by preeclampsia. Biol Res Pregnancy Perinatol 1985;6:121 – 5. [19] Evans RW, Powers RW, Ness RB, Cropcho LJ, Daftary AR, Harger GF, et al. Maternal and fetal amino acid concentrations and fetal outcomes during preeclamsia. Reproduction 2003; 125:785 – 90.
Plasma and urinary free amino acid concentrations in preeclamptic women in northern Nigeria Robert H. Glew a, George Melah b, Aliyu I. El-Nafaty b, Yekaterina Brandt c, D. Morris c, Dorothy J. VanderJagt a,* a
Department of Biochemistry and Molecular Biology, MSC08 4670, 1 University of New Mexico School of Medicine, Albuquerque, NM 87131-0001, USA b Department of Obstetrics and Gynaecology, Federal Medical Centre, Gombe, Gombe, Nigeria c Department of Surgery, University of New Mexico School of Medicine, Albuquerque, NM 87131-0001, USA Received 18 November 2003; received in revised form 29 December 2003; accepted 31 December 2003
Abstract Background: Preeclampsia/eclampsia is one of the major causes of maternal and fetal mortality in Nigeria and many other countries in sub-Saharan Africa. We determined if serum and urine concentrations of amino acids were abnormal in women with this disorder of pregnancy in Gombe, Nigeria. Methods: Free amino acids were measured in serum and urine of women (ages, 15 to 40 years) with preeclampsia/eclampsia (n = 37) and their pregnant age and gestational age matched controls (n = 16). Results: The concentrations of 19 of the 20 serum amino acids that are common in proteins were not significantly different between the control and preeclamptic groups. Phenylalanine was increased in the preeclamptic women compared to the controls ( p < 0.05); however, the difference between the two values was small (119 vs. 104 umol/l, respectively). The concentrations of urine free amino acids were indexed to urine creatinine (i.e., nmol/mg creatinine). Five amino acids were significantly elevated in the urine of preeclamptic patients vs. the control group: proline (100%, p < 0.05), the branched amino acids, valine (80%, p < 0.05), leucine (61%, p < 0.05), isoleucine (49%, p = 0.01), and methionine (52%, p < 0.005). On the other hand, the concentrations of a nutritionally non-essential amino acid, glycine ( p = 0.01), and an essential amino acid, histidine ( p = 0.01), were both reduced by about one-third in the urine of the preeclamptic women. Conclusions: Our findings indicate that there are disturbances in urinary amino acid excretion but not in the serum amino acid profile in the preeclamptic patients. The degree of aminoaciduria is probably not sufficient to significantly affect overall nitrogen balance or precipitate a deficiency of one or more of the essential amino acids. D 2004 Elsevier B.V. All rights reserved. Keywords: Preeclampsia; Hypertension; Proteinuria; Aminoaciduria; Pregnancy; Nigeria
1. Introduction
* Corresponding author. Tel.: +1-505-272-5799; fax: +1-505272-3518. E-mail address: [email protected] (D.J. VanderJagt). 0009-8981/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.cccn.2003.12.028
Preeclampsia/eclampsia is a major cause of maternal and perinatal mortality in northern Nigeria [1,2] and many other regions of sub-Saharan Africa. It is the leading cause of maternal mortality at the Specialist Hospital in Maiduguri, Nigeria [3]. Although the risk of a pregnancy developing into eclampsia can be
180
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
reduced by good antenatal care, there is still no method of preventing preeclampsia. Furthermore, although there is a growing literature pointing to diffuse endothelial dysfunction in preeclampsia/eclampsia [4], the cause of this multisystem disorder of pregnancy remains elusive. Preeclampsia usually occurs who have been normotensive after the 20th week of pregnancy and is characterized by hypertension, with or without edema, and proteinuria. The disturbances associated with preeclampsia include: hyperlipidemia, particularly hypertriglyceridemia [5,6]; excessive lipid peroxidation and insulin resistance [7]; and imbalances in the placental production of thromboxane A2 and prostacyclin [8]. There is a lack of consensus in the literature regarding amino acid concentrations in the blood of preeclamptic women. Li et al. [9] reported finding increased concentrations of aromatic amino acids, threonine, arginine, glycine, cystine and glutamic acid in the plasma of preeclamptic women. On the other hand, Kremenski et al. [10] found preeclampsia to be associated with increases in the plasma concentrations of aspartic acid, glutamic acid, threonine, glycine and asparagines, but not the aromatic amino acids or cystine. Contradicting the findings, D’Aniello et al. [11] reported a five-fold lower mean plasma arginine concentration in preeclamptic women relative to controls matched for age and stage of gestation. Moreover, these investigators found that the concentrations of most of the other 19 common amino acids were significantly reduced, on average by about 50%, in the serum of preeclamptic subjects relative to the controls. The present study was conducted in order to determine if there were abnormalities in the concentrations of the serum amino acids of preeclamptic women in northern Nigeria. Our hypothesis was that the proteinuria-based protein wastage associated with preeclampsia, together with the marginal protein intakes of the general population in northern Nigeria where this study was conducted, would cause the preeclamptic women to have reduced serum concentrations of one or more of the essential amino acids. We also took advantage of the opportunity to compare the urinary excretion of free amino acids between preeclamptic women and healthy pregnant controls.
2. Materials and methods 2.1. Subjects Pregnant women were recruited into the study at the Federal Medical Centre, Gombe and the Specialist Hospital, Gombe, Nigeria. The diagnosis of preeclampsia was made in 37 pregnant women based on the criteria proposed by the American College of Obstetricians and Gynecologists which included: cerebral or visual disturbances, epigastric pain, pulmonary edema or cyanosis, a systolic blood pressure z 140 mm Hg or a diastolic blood pressure z 90 mm Hg and proteinuria. We set the cut-off for proteinuria at 190 mg total protein/g creatinine [12]. Protein was estimated using the method of Bradford [13] with albumin serving as the protein standard. Serum uric acid concentrations were not determined. Ten milliliters of venous blood was collected into clean, dry vacutainers and allowed to clot for 45 min at room temperature. After centrifugation at 5000 g for 8 min, the clear supernatant was aliquoted into 2-ml cryovials and stored for 2– 6 weeks at 43 jC until which time the specimens were transported in the frozen state to Albuquerque for analysis. At the same time blood was drawn, urine was collected into sterile plastic cups, aliquoted into cryovials and stored and transported as described above for blood serum. The control population consisted of 16 healthy pregnant women whose mean age and gestational age matched the corresponding mean values of the preeclamptic subjects. The study was approved by the Human Ethics Committee at the Federal Medical Centre, Gombe and the University of New Mexico School of Medicine, and informed consent was obtained for all subjects. 2.2. Amino acid analysis A 50-Al aliquot of serum or urine was mixed with an equal volume of 0.5 mmol/l norleucine, the internal standard. Norleucine was used as the internal standard because it is not commonly found in proteins. For the purpose of deproteination, the mixture was pipetted into an ultrafiltration device (Ultrafree-MC, Milllipore, Bedford, MA) and centrifuged at 2400 g for 30 min. No special precautions were taken with regard to tryptophan. An aliquot of the supernatant was transferred into a 6 50-mm glass tube and dried in a
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
vacuum. Following drying, 10 Al of redry solution (methanol/ 1.0 mol/l sodium acetate/ triethylamine, 2:2:1) was added after which the sample was redried. Next, 20 Al of phenylisothiocyanate (PITC) reagent (methanol/water/triethylamine/phenylisothiocyanate, 7:1:1:1) was added under alkaline conditions. The reaction was allowed to proceed for 20 min at 20 jC to produce phenylthiocarbamoyl amino acids (PTCAA) [14]. Excess reagent was removed at 20 jC with the aid of a vacuum pump. After drying the sample, the PTC-AA were dissolved in 0.1 ml buffer A (0.07 mol/ l sodium acetate, pH 6.5 containing 2.5% (v/v) acetonitrile and 1 ppm EDTA, Ca+ 2 salt). The derivatized amino acids were then separated by reverse-phase high performance liquid chromatography using an Ultra WISP 715 autoinjector and a Waters Pico Tag solvent delivery system (Model 510). Elution of PTC-AA was achieved using gradient consisting of buffer A and buffer B (acetonitrile/water/methanol, 45:40:15, by volume) over 64 min at a flow rate of 1 ml/min [15]. The column effluent was monitored at 254 nm and the signal was integrated using Millenium 1999 ver. 3.20 (Waters, Milford, MA) software. The retention times of the various amino acids were determined using an amino acid calibration mixture. Quantitation of individual amino acids was achieved by monitoring the absorption of the column eluate at 254 nm and comparing the areas under the individual peaks with those of the corresponding amino acid standards. The average standard error of the mean for the 20 amino acids was 3.1%.
181
Table 1 Comparison of characteristics of preeclamptic/eclamptic women and their matched healthy controls from Gombe, Nigeria Parameter
Matched controls, mean F S.D. (n = 16)
Age (years) BP, systolic (mm Hg) BP, diastolic (mm Hg) Parity Gestational age (weeks) Urine protein/ creatinine ratio (mg/g)
26.9 (7.2) 107 (7) 69 (8) 2.5 (2.8) 35.3 (5.5) 121 (44)
Preeclampsia/ eclampsia, mean F S.D. (n = 37)
p-value
25.0 (7.3) 151 (25)
NS < 0.001
103 (14)
< 0.001
2.3 (2.9) 35.4 (3.9) 1115 (2308)
NS NS < 0.001
NS, non-significant ( p>0.05).
significant differences in the two groups in terms of parity (2.3 –2.5) or gestational age (approx. 35 weeks). The mean diastolic and systolic blood pressures of the preeclamptic women and their total urine total protein excretion (expressed as mg protein/g creatinine) values were significantly increased relative to the pregnant controls, which is expected given that blood pressure and urinary protein concentrations were the factors distinguishing the two populations. The mean gestational age of both control and preeclamptic subjects was approximately 35 weeks. The controls were selected to match the gestational age of the preeclamptic patients.
2.3. Statistical methods
3.2. Serum amino acids
The two-sample t-test was used to compare variables between patients with preeclampsia and controls using the NCSS 2001 Statistical System for Windows (Kaysville, UT). A p-value < 0.05 was considered significant.
Table 2 shows that, with the exception of phenylalanine, the serum concentrations of the common amino acids were not significantly different between the pregnant controls and the subjects with preeclampsia. Furthermore, even though the phenylalanine concentration was significantly higher in the preeclamptic women compared to the controls, the difference between the two values was < 20%. The serum total amino acid concentration was not significantly different between the control and preeclamptic populations (3030 vs. 3150 Amol/l). Noteworthy was our finding that, in contrast to the study by D’Aniello et al. [11] which reported a five-fold reduction in the plasma arginine concentration of preeclamptic women, we
3. Results 3.1. Comments on the subjects As shown in Table 1, the mean ages of the preeclamptic women and the pregnant controls (25.0 and 26.9 y) were not significantly different, nor were there
182
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
Table 2 Comparison of the concentrations of amino acids in the serum of preeclamptic/eclamptic women and matched controls (Amol/l) Amino acid
Matched controls (n = 16), mean F S.D.
Preeclampsia/eclampsia (n = 37), mean F S.D.
Asp Glu Ser Asn Gly Gln His Thr Ala Arg Pro Tyr Val Met Cys Ile Leu Phe Trp Lys Total
62 (18) 236 (62) 214 (46) 114 (22) 278 (62) 438 (101) 47 (25) 92 (32) 438 (110) 148 (110) 217 (64) 57 (12) 173 (43) 24 (5) 8 (3) 55 (15) 175 (34) 104 (23) 7 (6) 141 (54) 3030
63 (15) 242 (101) 218 (40) 107 (28) 308 (78) 455 (115) 48 (23) 107 (44) 451 (121) 151 (87) 221 (62) 59 (16) 180 (36) 25 (7) 8 (4) 54 (16) 175 (41) 119 (25)a 8 (8) 147 (46) 3150
a
p < 0.05.
found the serum arginine concentrations of the pregnant controls and preeclamptic women to be essentially the same. 3.3. Urine amino acids The urine concentrations of amino acids were expressed on the basis of urine creatinine (i.e., nmol/mg creatinine). In contrast to serum where the free amino acid profiles there were nearly indistinguishable, a number of relatively large and statistically significant differences were found when we compared the concentrations of the 20 common amino acids in the urine of the preeclamptic and control pregnant women. The concentrations of two nutritionally non-essential amino acids, namely glycine and histidine, were both reduced by about one-third in the urine of the preeclamptic women. On the other hand, the following five amino acids were present in increased amounts in the urine of the preeclamptic subjects relative to the controls: proline (100%, p < 0.05); the branched amino acids, valine (80%,
p < 0.0.05), leucine (61%, p < 0.0.05) and isoleucine (49%, p = 0.01); and methionine (52%, p < 0.005). No correlations were found when the urine concentration (indexed to creatinine) of proline, valine, leucine, isoleucine or methionine was plotted against the total urinary protein, also indexed to creatinine.
4. Discussion The present study yielded two main findings. First, the serum amino acid concentration profiles of preeclamptic women in Gombe, Nigeria were not very different from those of healthy pregnant women matched for age, gestational age and parity. Second, preeclampsia was associated with abnormalities in amino acid excretion. The type of aminoaciduria we documented in the preeclamptic women in our study was most likely of the secondary kind, as opposed to the overflow aminoaciduria one sees in fulminant hepatic failure [16]. It is noteworthy that four of the five amino acids that were present in increased amounts in the urine of the preeclamptic women are regarded collectively as belonging to the hydrophobic amino acid class. Glycine and the essential amino acid histidine were excreted in significantly reduced quantities relative to the controls. How might one account for the altered excretion of certain amino acids in preeclampsia? The excretion of greater than normal quantities of amino acids in the urine could be accounted for by a compromise of the function of one or more of the amino acid transporters in the proximal convoluted tubule of the kidney. There are five stereospecific reabsorptive systems for amino acids [17]: Group I, neutral monoamino, monocarboxylic and cyclic amino acids; Group II, cystine and dibasic amino acids; Group III, imino acids and glycine; Group IV, dicarboxylic amino acids; and Group V, h-amino acids. Most amino acids are reabsorbed by more than one transporter. Since methionine excretion was increased and histidine excretion was decreased in the urine of the preeclamptic subjects, and because both of these amino acids fall into the Group I class of transporters, a defect in one of the amino acid transporters in the kidney in the preeclamptic women in our study cannot account for the selective aminoaciduria we observed in women with this hypertensive disorder of pregnancy. By similar reasoning, since glycine and
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
proline are both Group III amino acids, our finding of decreased glycine excretion and increased proline excretion in the urine of the preeclamptic women relative to the healthy pregnant women argues against a hypothesis asserting that defects in one or more of the amino acid transporters in the kidney could account for the abnormal quantities of particular amino acids we observed in the urine of the preeclamptic women in Gombe. We have no alternative explanation for our amino acid findings. The results of the present study of urine concentrations of amino acids in preeclamptic women in northern Nigeria are not in accord with the results of a similar study by Velco et al. [18]: whereas we found abnormalities in the urinary excretion of seven amino acids in 37 preeclamptic women, they found no differences in the urine concentrations (indexed to creatinine) of amino acids in four women with proteinuric hypertension when compared with that in normal pregnancy. Our results regarding serum amino acid concentrations in preeclampsia also differ markedly from D’Aniello and coworkers’ [11], who reported a five-fold reduction in the mean plasma arginine concentration of 12 preeclamptic women relative to an equal number of pregnant controls matched for age and gestational age. The mean serum arginine concentrations of our preeclamptic and control pregnant were essentially the same (Table 2). Since hemodilution and an increase in the glomerular filtration rate are associated with pregnancy and both are reduced during preeclampsia, we expected to find significant increases in the serum concentrations of many amino acids in the preeclamptic subjects. A reduction in hemodilution in the preeclamptic subjects would tend to elevate amino acid concentrations, as would a decrease in the glomerular filtration rate. However, because we did not determine the concentration of albumin in the sera of any of the subjects in our study, we cannot comment on the degree to which hemodilution might have been a factor in affecting serum amino acid values of the preeclamptic subjects. Similarly, because we did not estimate the glomerular filtration rate of the subjects in our study, we are not in a position to comment on how this particular aspect of kidney function might have influenced our findings with respect to serum and urine amino acid concentrations.
183
The fact that we did not find noteworthy or significant differences in the concentrations of amino acids in the serum of preeclamptic vs. healthy pregnant women leads us to conclude that abnormalities in maternal amino acid metabolism were not associated with preeclampsia in northern Nigeria. Furthermore, this observation also suggests that transplacental transfer of amino acids from the mother to the fetus was probably not affected in the preeclamptic Nigerian subjects in our study. In contrast to our results, Evans et al. [19] recently reported finding increased concentrations of a number of amino acids in the serum of preeclamptic subjects in the US which they attributed to enhanced placental amino acid transport or reduced fetal amino acid utilization. The discrepancy between their findings and ours with respect to serum amino acid concentrations in preeclampsia may be accounted for by the fact that the preeclamptic subjects in our study were probably significantly less well nourished than the hypertensive, proteinuric subjects in their study. However, our data do point to disturbances in urinary amino acid excretion in the preeclamptic women we studied. We can only speculate regarding the mechanism that might account for the abnormal Table 3 Comparison of the concentrations of essential amino acids in the urine of preeclamptic/eclamptic women and controls (nmol/mg Cr) Amino acid
Matched controls (n = 16), mean F S.D.
Preeclampsia/eclampsia (n = 37), mean F S.D.
p-value
Asp Glu Ser Asn Gly Gln His Thr Ala Arg Pro Tyr Val Met Cys Ile Leu Phe Trp Lys Total
51.4 (38.7) 248 (72.4) 598 (189) 358 (169) 3450 (1511) 774 (299) 1090 (399) 210 (73.2) 651 (254) 17.4 (9.0) 23.1 (9.2) 103 (41.6) 47.8 (12.8) 22.0 (5.7) 28.0 (11.5) 27.0 (10.3) 65.1 (26.4) 73.7 (26.8) 78.9 (43.2) 139 (253) 8055
43.2 (43.3) 286 (187) 540 (341) 333 (267) 2382 (1654) 597 (373) 760 (579) 194 (189) 584 (377) 26 (25.6) 45.7 (55.2) 94.3 (59.1) 85.3 (77.2) 31.3 (16.0) 39.1 (28.4) 39.7 (18.4) 105 (73.0) 94.6 (53.9) 89.4 (70.6) 81.5 (78.9) 6451
NS NS NS NS < 0.05 NS 0.01 NS NS NS < 0.05 NS < 0.05 0.01 NS 0.01 < 0.05 NS NS NS
184
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185
pattern of amino acids in the urine of these preeclamptic women. It is conceivable that the excessive protein in the urine of the preeclamptic women causes the uptake by proximal tubule cells of quantities of protein so large as to disturb the amino acid transport systems of these cells that are responsible for reabsorbing filtered free amino acids. However, since we did not carry out a thorough analysis of the nature or size of the proteins that were present in the urine of the preeclamptic women in our study, we cannot make a strong statement regarding the basis of the proteinuria they were experiencing. Much of the excessive urinary protein in the preeclamptic subjects in the present study can be explained by increased glomerular permeability. This deduction is based on the fact that in the 37 preeclamptic women, albumin accounted for 30– 70% of the total protein. It is well known that in glomerular proteinuria, the most abundant protein in urine is albumin [16]. However, since albumin did not account for all of the urinary protein in the preeclamptic women in our study, we suggest that defective tubular reabsorption also contributed significantly to the proteinuria they exhibited. In tubular proteinuria, albumin excretion is only slightly increased and proteins other than and of lower molecular weight than albumin usually account for most of the protein found in the urine [16]. We suggest that the mixture of albumin and non-albumin proteins in the urine of the preeclamptic women in the present study points to the co-presence of both glomerular and tubular proteinuria in these women. Finally, there is the question of what the detrimental consequence(s) might be of the increased excretion of certain amino acids by the preeclamptic subjects in the present study. Although the increased quantities of proline, valine, leucine, isoleucine and methionine excreted in the urine of the preeclamptic women were considerable, they would probably not be sufficient to significantly affect overall nitrogen balance or precipitate a deficiency of one or more of the essential amino acids. Since insulin plays such a key role in regulating the metabolism of the branched-chain amino acids, the fact that the quantities of valine, leucine and isoleucine were each increased by >50% in the urine of the preeclamptic subjects (Table 3) raises questions regarding insulin status and function in the preeclamptic women in our study.
In conclusion, although we have documented significant differences in the quantities of seven different amino acids in the urine of preeclamptic women, in light of the fact that the mean serum amino acid profile for this group of hypertensive, proteinuric subjects was, overall, so similar to that of the pregnant controls and because the degree of aminoaciduria in the patient group was only moderate, it is unlikely that differences in the urinary excretion of amino acids are of major clinical significance in the preeclamptic women we studied. Acknowledgements This study was supported by a Minority International Research Training (MIRT) grant from the Fogarty International Center of the National Institutes of Health. References [1] Harrison KA. Child-bearing, health and social priorities: a survey of 22 774 consecutive hospital births in Zaria, Northern Nigeria. Br J Obstet Gynaecol 1985;S5:1 – 119. [2] El-Nafaty AU. Pre-clampsia and eclampsia: a clinical guide for early detection and management for health workers in developing countries. Maiduguri: Ed-Linform Services; 1998. [3] Bobzom DN, Muna M. Maternal mortality at the State Specialist Hospital, Maiduguri, Borno State. Annals of Borno 1994;11/12:270 – 4. [4] Roberts JM, Redman CW. Preeclampsia: more than pregnancy-induced hypertension. Lancet 1993;341:1447 – 51. [5] Kaaja R, Tikkanen MJ, Viinikka L, Ylikorkala O. Serum lipoproteins, insulin and urinary prostanoid metabolites in normal and hypertensive pregnant women. Obstet Gynecol 1995; 85:353 – 6. [6] Gratacos E, Casals E, Sanllehy C, Cararach V, Alanos PL, Fortuny A. Variation in lipid levels during pregnancy in women with different types of hypertension. Acta Obstet Gynecol Scand 1996;75:896 – 901. [7] Walsh SW. Lipid peroxidation in pregnancy. Hypertens Pregnancy 1994;13:1 – 32. [8] Walsh SW, Behr MJ, Allen NH. Placental prostacyclin production in normal and toxemic pregnancies. Am J Obstet Gynecol 1985;151:110 – 5. [9] Li J, Sun J, Lu H. Determination of serum free amino acids and ammonia in pregnant women with severe pregnancy induced hypertension. Zhonghua Fu Chan Ke Zhi 1996;31: 468 – 70. [10] Kremenski I, Borison I, Barov D, Katsulov A. The plasma amino acid profile of women with a normal pregnancy and in pre-eclampsia. Akush Ginekol (Sofia) 1990;29:5 – 9.
R.H. Glew et al. / Clinica Chimica Acta 342 (2004) 179–185 [11] D’Aniello G, Tolino A, Fisher G. Plasma L-arginine is markedly reduced in pregnant women affected by preeclampsia. J Chromatogr 2001;753:427 – 31. [12] Barratt M. Proteinuria. Br Med J 1983;287:1489 – 90. [13] Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein using the principle of protein dye-binding. Anal Biochem 1976;72:248 – 54. [14] Cohen SA, Strydom DJ. Amino acid analysis utilizing phenylisothiocyanate derivatives. Anal Biochem 1988;174:1 – 16. [15] Buzzigoli G, Lanzone L, Ciorciaro D, Fascerra S, Cerri M, Scandroglio A, et al. Characterization of a reversed-phase high performance liquid chromatographic system for determination of blood amino acids. J Chromatogr 1990;507: 85 – 93.
185
[16] Johnson AM, Rohlfs EM, Silverman L. Proteins. In: Burtis CA, Ashwood ER, editors. Tietz textbook of clinical chemistry. Philadelphia: W.B. Saunders; 1999. p. 477 – 540. Chapter 20. [17] Bonnardeaux A, Bichet DG. Inherited disorders of the renal proximal tubule. sixth edition. Brenner and Rector’s the kidney, vol. 2. Philadelphia: W.B. Saunders; 2000. p. 1656 – 98. [18] Velco CJ, Reynolds JW, Hawkins DF, Meek JH. Renal clearance of amino acids in pregnancy complicated by preeclampsia. Biol Res Pregnancy Perinatol 1985;6:121 – 5. [19] Evans RW, Powers RW, Ness RB, Cropcho LJ, Daftary AR, Harger GF, et al. Maternal and fetal amino acid concentrations and fetal outcomes during preeclamsia. Reproduction 2003; 125:785 – 90.