0021-972X/99/$03.00/0 The Journal of Clinical Endocrinology & Metabolism Copyright © 1999 by The Endocrine Society
Vol. 84, No. 3 Printed in U.S.A.
Long-Term Treatment with Growth Hormone Has No Persisting Effect on Lipoprotein(a) in Patients with Turner’s Syndrome ¨ PPER, A. GRADEHAND, P. KIENCKE, F. WAHN, U. QUERFELD, S. DO H.-J. ZEISEL
Departments of Pediatrics and Medical Statistics (P.K.), University of Cologne, 50924 Cologne, Germany; and Clinical Research Institute (H.-J.Z.), 79232 March, Germany ABSTRACT Treatment with recombinant human GH (rhGH), alone or in combination with the anabolic steroid oxandrolone (OX), has been recommended for girls with Turner’s syndrome to improve final height. Several cardiovascular risk factors have been described in patients with Turner’s syndrome, but the effect of therapy with rhGH and OX on lipoprotein(a) [Lp(a)] has not been investigated. Lp(a) serum levels and apolipoprotein(a) phenotypes were determined in 46 girls with Turner’s syndrome (aged 6 –15 yr) during treatment with different
combinations of rhGH and OX for 24 –36 months (median, 27 months). Lp(a) serum levels showed little variation during 30 months of treatment in all treatment groups. Lp(a) levels showed no significant change in 25 patients receiving only rhGH and in 21 patients receiving rhGH and OX in combination. Treatment effects were independent of apolipoprotein(a) phenotypes and were not influenced by pubertal status. These data indicate that long term administration of rhGH has no significant impact on serum Lp(a) levels in girls with Turner’s syndrome. (J Clin Endocrinol Metab 84: 967–970, 1999)
HORT STATURE is the most common clinical feature of Turner’s syndrome (1). Treatment with recombinant human GH (rhGH) is effective in increasing the final height of patients with Turner’s syndrome (1, 2). Therefore, although there is no deficiency of GH in this disease, therapy with rhGH has been recommended alone or in combination with anabolic agents such as oxandrolone (OX) as standard therapy when the height is below the fifth percentile of the normal female growth curve (1). In addition, patients frequently receive estrogen therapy (or combined estrogens and gestagens) to promote the development of female sexual appearance. Turner’s syndrome is also associated with clustering of cardiovascular risk factors, i.e. hypertension, obesity, insulin resistance, and an increased incidence of type II diabetes (3, 4). Although the synergistic effect of multiple risk factors on atherosclerosis has been established in large population-based studies (5) and in autopsy studies in young children and adults (6), there is surprisingly little information on the effect of hormonal treatment on these cardiovascular risk factors, although both rhGH and estrogens are now used in many patients with Turner’s syndrome for prolonged periods. Of special interest is the effect of these hormones on lipoprotein(a) [Lp(a)]. Lp(a) is an independent risk factor for atherosclerosis and thrombosis (7) and seems to be of greatest clinical importance in young patients with cardiovascular disease (8, 9). Plasma Lp(a) levels vary over 1000-fold between individuals and are largely determined by the size of the distinguishing apoplipoprotein(a) [apo(a)], which shows enormous size heterogeneity (10).
Epidemiological studies have demonstrated an inverse relationship between the size of apo(a) and the plasma level of Lp(a), i.e. small apo(a) phenotypes are associated with high levels of Lp(a), and large apo(a) phenotypes are associated with low levels of Lp(a) (7, 11). Although plasma Lp(a) levels are for the most part genetically determined, the usually stable intraindividual Lp(a) levels show considerable variations during hormonal treatment. Thus, rhGH has been shown in most (12–17), but not all (18 –21), studies to increase plasma Lp(a), whereas treatment with estrogens, gestagens, testosterone, or anabolic steroids decreases Lp(a) levels (22, 23). Studies evaluating the effects of rhGH treatment have been mainly conducted in GH-deficient adults and children, and there are no reports of long term effects of this form of hormonal therapy in patients with Turner’s syndrome. We therefore studied serum Lp(a) levels and apo(a) phenotypes in patients with Turner’s syndrome treated with different combinations of rhGH and OX. Subjects and Methods A total of 46 girls with Turner’s syndrome (age, 6 –15 yr; median age, 11 yr) participating in a multicenter trial in Germany were studied. The results of this trial and details of the study population have been published previously (24, 25). The study was designed to evaluate the effect of rhGH with or without additional treatment with OX on height gain. Five patients had a mosaic form of Turner’s syndrome (45,X; 46,XX) and all other girls had the 45,X phenotype. Patients were randomized to 1 of 4 groups and were studied for a total of 24 –36 months (median, 27 months; range, 24 – 45 months; only 1 patient .36 months). The 4 groups of patients varied in dosage of rhGH and OX and the timing of combination of rhGH with OX (Table 1). Before randomization, parental/guardial informed consent was obtained for every patient. During the study, five patients received estrogen treatment, and another five patients received thyroid hormone substitution. All patients were studied before treatment and followed every 3
Received March 20, 1998. Revision received October 29, 1998. Accepted November 16, 1998. Address all correspondence and requests for reprints to: Uwe Querfeld, M.D., University Children’s Hospital, Joseph-Stelzmann-Strasse 9, 50924 Cologne, Germany. E-mail: [email protected]
JCE & M • 1999 Vol 84 • No 3
QUERFELD ET AL.
TABLE 1. Patients and treatment modalities Group
GH: 18 IU/m2 z week GH: 18 IU/m2 z week OX: 0.1 mg/kg z day GH: 18 IU/m2 z week
GH: 24 IU/m2 z week GH: 18 IU/m2 z week OX: 0.05 mg/kg z day GH: 24 IU/m2 z week
GH: 18 IU/m2 z week OX: 0.1 mg/kg z day
GH: 18 IU/m2 z week OX: 0.05 mg/kg z day
GH: 24 IU/m2 z week GH: 18 IU/m2 z week OX: 0.05 mg/kg z day GH: 24 IU/m2 z week OX: 0.05 mg/kg z day GH: 24 IU/m2 z week
months thereafter. Blood was drawn at each visit in the morning in the fasting state; an aliquot of serum was immediately frozen and shipped on dry ice to a central reference laboratory (laboratory of H.-J.Z.). Upon arrival, these samples were thawed and either used for analysis (of anti-GH and antihost cell antibodies) or immediately divided into aliquots and stored at 270 C in special aluminum cap-sealed small glass tubes for long term storage. Lp(a) and apo(a) phenotypes were measured in aliquots that had been stored without thawing for 5–9 yr (median storage time, 6.5 yr). Serum Lp(a) concentrations (at 3-month intervals) were measured with a one-step enzyme-linked immunosorbent assay (Immuno, Heidelberg, Germany). All samples were assayed in quadruplicate. The apo(a) phenotype was determined in every patient with a commercially available SDS-PAGE system (PHAST system, Pharmacia, Freiburg, Germany) using a 4 –15% gradient gel followed by Western blotting with a polyclonal sheep anti-apo(a) antibody (Immuno). Apo(a) phenotypes were designated as suggested by Utermann et al. (26); if double band phenotypes were present (heterozygotes), the apo(a) phenotype was determined according to the predominant band.
Statistical analysis To accommodate for the statistical problem of repeated measurements in a single patient before and after treatment, analysis of covariance with repeated measures was conducted using BMDP2V (release 7.0) software and incorporating the pretreatment Lp(a) level as a covariate, as suggested by Frison and Pocock (27) and others (28). The purpose of the study, its protocol, and the documents describing the study medication were first approved by the ethical committee of the University of Hamburg, Germany, and thereafter by the ethical committees of all other participating hospitals. The German Health Authority was notified of this study, and the relevant documentation was submitted. This study was performed according to good clinical practice guidelines and was designed to support the registration of rhGH for the indication short stature due to Turner’s syndrome, which has been granted in 1992 by the European Health authorities. Sera taken during the course of the study were stored in the laboratory of one of the authors (H.-J.Z.) and analyzed for Lp(a) and apo(a) phenotypes without any financial or other support from the manufacturer of the study medication.
Results Lp(a) serum levels and effect of rhGH
Lp(a) serum levels showed very little variation over 30 months of treatment when all patients were analyzed either combined (Fig. 1) or separately within treatment groups (Table 2). Wider fluctuations were seen after 30 months of treatment; however, this was most likely due to the small number of patients available at this period of treatment (n 5 9 at 33 months; n 5 5 at 36 months; data not shown). Only one patient showed a consistent increase in serum Lp(a) levels [from 18 mg/dL before to 66 mg/dL after 30 months of rhGH treatment; apo(a) phenotype S4/S1; group 2]. Two patients had a serum Lp(a) level of more than 30 mg/dL at baseline, and six other patients had levels above 30 mg/dL at any time point during the study; all of these patients had small apo(a) phenotypes. To investigate whether rhGH had a significant effect on Lp(a), serum levels over 24 months after start of treatment
were compared to baseline levels in all patients receiving only rhGH during the first 2 yr (groups 1 and 3; n 5 25). Lp(a) levels showed no significant change over time (P 5 0.1186). Effects of rhGH and OX on Lp(a)
To study the combined effects of rhGH and OX on Lp(a), serum levels after the start of treatment were compared to baseline levels in all patients receiving rhGH and OX combined (group 2 over 3 yr and group 4 over the first 2 yr; n 5 21). Lp(a) levels showed no significant change over time (P 5 0.1927). Effect of apo(a) phenotypes within treatment groups
For easier statistical comparison, we grouped the apo(a) phenotypes of all patients into low mol wt (LMW) phenotypes (phenotypes F, B, S1, and S2) and high mol wt (HMW) phenotypes (all phenotypes above S2). As expected, Lp(a) levels were lower in patients with HMW phenotypes (median, 0.9 mg/dL at baseline; n 5 27) than in those patients with LMW phenotypes (median, 13.2 mg/dL at baseline; n 5 16). We then studied the effect of rhGH (25 patients in groups 1 and 3) in patients with HMW and LMW apo(a) phenotypes. Treatment with rhGH had no significant effect in patients with either HMW (n 5 14; P 5 0.2882) or LMW (n 5 8; P 5 0.3016) phenotypes (3 patients had no identifiable phenotype). In patients receiving combined treatment with rhGH and OX, Lp(a) levels showed no significant change in either HMW (n 5 13; P 5 0.2082) or LMW (n 5 7; P 5 0.3776) phenotypes (one patient without identifiable phenotype). Effect of puberty on Lp(a) within treatment groups
We then studied whether Lp(a) levels would be influenced by puberty by analyzing Lp(a) in those patients receiving rhGH who entered puberty during the study (n 5 15). There was no significant change in Lp(a) in either patients entering puberty (n 5 15; P 5 0.5488) or in prepubertal patients (n 5 10; P 5 0.3038). Similarly, in patients receiving rhGH and OX in combination, Lp(a) showed no significant change in either patients entering puberty (n 5 13; P 5 0.1033) or in prepubertal patients (n 5 8; P 5 0.3081). Discussion
The present study was conducted to evaluate whether long term treatment with rhGH would lead to a persistent increase in Lp(a) serum levels, which would amount to a serious disadvantage of this form of treatment, especially if one considers the presence of other risk factors for cardiovascular disease in patients with Turner’s syndrome (3, 4). Overall, the results
GH AND Lp(a) IN TURNER’S SYNDROME
FIG. 1. Lp(a) levels in 46 patients with Turner’s syndrome.
TABLE 2. Serum Lp(a) levels within treatment groups Group
Median D(mg/dL) n
Median D(mg/dL) n
Median D(mg/dL) n
Median D(mg/dL) n
16 2.40 16 1.55 5
1.00 0.10 8
1.70 0.80 9
2.20 1.30 9
1.30 0.40 7
1.00 0.10 8
1.15 0.25 8
0.75 20.15 8
0.85 20.05 8
0.50 20.40 5
1.60 0.70 1
0.40 21.45 16
0.20 21.65 14
0.40 21.45 14
0.50 21.35 14
0.50 21.35 16
0.50 21.35 15
1.50 20.35 15
0.50 21.35 13
0.85 21.00 10
2.15 0.30 7
6.00 3.60 14
4.00 1.60 16
2.50 0.10 15
5.10 2.70 13
5.40 3.00 13
2.55 0.15 14
1.70 20.70 15
1.50 20.90 13
1.70 20.70 13
1.55 20.85 14
1.50 20.05 4
1.30 20.25 5
0.80 20.75 3
3.10 1.55 4
3.00 1.45 5
4.80 3.25 5
1.60 0.05 5
4.70 3.15 5
3.15 1.60 4
4.00 2.45 3
indicate that changes were small and, if present, only of transient nature. Treatment effects were independent of apo(a) phenotypes and were not influenced by pubertal status. The median Lp(a) levels showed very little fluctuation during rhGH therapy; moreover, intraindividual fluctuations were also small, and there was only a single patient who showed a consistent increase in Lp(a) levels. This is in contradiction to many studies evaluating the effects on rhGH on Lp(a) (12–17). However, to our knowledge there have been 4 studies published to date showing no effect of rhGH on Lp(a); this includes a double blind, placebocontrolled study of 18 adults with rhGH deficiency treated for 2 months (18), an uncontrolled study of 27 children with GH deficiency treated for 12 months (19), a study of 18 adults with GH deficiency or hypopituitarism treated in a double blind, placebo-controlled manner for 6 months and in an open study for an additional 6 months (20), and an uncontrolled study of 12 children with b-thalassemia treated for 24 –36 months (21). Interestingly, a transient elevation of Lp(a) was found in the latter study, with a return to baseline after 36 months. We can only summarize the available data by stating that there is conflicting evidence at present regarding the short term effects of rhGH on serum Lp(a) levels, whereas the only published long
term study (21) as well as our own data argue against a persisting effect of rhGH on Lp(a). To our knowledge, this is the first study examining long term effects of rhGH in a sizable number of subjects with Turner’s syndrome. In two small previous studies in patients with Turner’s syndrome, a significant rise in serum Lp(a) levels was found in seven girls after 6 –9 months (16), and no effect on Lp(a) was observed in six girls after 6 months, respectively, of rhGH treatment (29). It should be noted that the present study does not include male subjects. Most rhGH treatment studies, regardless of whether an increase in Lp(a) was observed, have included both men and women, but little information is provided on differences in the effect of rhGH on Lp(a) between males and females. In two studies, the increase in Lp(a) was much less pronounced in females (14, 30); therefore, gender differences in the response to rhGH cannot be ruled out (17). Failure to detect a significant effect of rhGH on Lp(a) could be due to ethnic differences between populations, which have a profound impact on Lp(a) (31); however, the studies documenting no effect of rhGH treatment on Lp(a) have been performed on English adults (18, 20), Japanese children (19), and Chinese children (21), making this explanation unlikely.
QUERFELD ET AL.
One of the reasons for not detecting a change in Lp(a) levels might be the presence of very low Lp(a) levels, e.g. in patients with HMW apo(a) phenotypes. Due to the vast size polymorphism of apo(a), it is almost impossible to design a study large enough to compare individuals with identical apo(a) phenotypes; many clinical studies have therefore compared patients with LMW and HMW phenotypes, with a cut-off at phenotype S2 (26, 32). We therefore also analyzed Lp(a) levels with respect to apo(a) phenotypes, but again found only very inconsistent small changes. In most studies observing an effect of rhGH on Lp(a), this was independent of the apo(a) phenotype (14, 15, 30, 33). In view of the well known decline in detectable Lp(a) levels with time of storage (34–36) and especially with repeated freezing and thawing (37), it could also be argued that we have failed to detect a major effect of rhGH on Lp(a) because we studied frozen sera. However, storage effects were minimized by the precautions taken of freezing sera in special containers without a break in the cooling chain. In addition, a major effect of long term storage seems unlikely for several reasons. We did detect some within-patient variations in Lp(a) (independent of storage time) and, as expected, major differences in Lp(a) between patients, depending on the individual apo(a) phenotype. However, although it has been stated that storage mainly affects the LMW isoforms (34), treatment effects were not significant in either LMW or HMW isoforms. Moreover, a strong effect of storage would have resulted in higher Lp(a) levels in the more recent samples (drawn near the end of the study) and, if anything, would have resulted in an overestimation of the effect of GH treatment. In conclusion, the present study shows that long term treatment with rhGH had no significant persisting effect on serum Lp(a) levels in girls with Turner’s syndrome. As these patients are known to have several risk factors for cardiovascular disease, the absence of a lasting effect on Lp(a) is reassuring. At present, however, the mechanisms of hormonal effects on Lp(a) remain largely unknown. Acknowledgments We thank the organizers of the German multicenter study, Prof. Dr. N. Stahnke, University Children’s Hospital Hamburg, and Prof. Dr. E. Keller, University Children’s Hospital Leipzig, for the permission to analyze the sera of patients from this study.
References 1. Saenger P. 1996 Turner’s syndrome. N Engl J Med. 335:1749 –1754. 2. van den Broeck J, Massa GG, Attanasio A, et al. 1995 Final height after long-term growth hormone treatment in Turner syndrome. J Pediatr. 127:729 –735. 3. Ross JL, Feuillan P, Long LM, Kowal K, Kushner H, Cutler GBJ. 1994 Lipid abnormalities in Turner syndrome. J Pediatr. 126:242–245. 4. Gravholt CH, Naerra RW, Nyholm B, et al. 1998 Glucose metabolism, lipid metabolism, and cardiovascular risk factors in adult Turner’s syndrome. Diabetes Care. 21:1062–1070. 5. Wilson PW, Evans JC. 1993 Coronary artery disease prediction. Am J Hypertens. 6:309S–313S. 6. Berenson GS, Srinivasan SR, Bao W, Newman III WP, Tracy RE, Wattigney WA. 1998 Assocation between multiple cardiovascular risk factors and atherosclerosis in children and young adults. N Engl J Med. 338:1656. 7. Scanu A, Fless GM. 1990 Lipoprotein(a). Heterogeneity and biological relevance. J Clin Invest. 85:1709 –1715. 8. Dahlen GH, Guyton JR, Attar M, Farmer JA, Kautz JA, Gotto AMJ. 1986 Association of levels of lipoprotein Lp(a), plasma lipids, and other lipoproteins with coronary artery disease documented by coronary angiography. Circulation. 74:758 –765.
JCE & M • 1999 Vol 84 • No 3
9. Sandkamp M, Funke H, Schulte H, Ko¨hler E, Assmann G. 1990 Lipoprotein(a) is an independent risk factor for myocardial infarction at a young age. Clin Chem. 36:20 –23. 10. Lackner C, Cohen JC, Hobbs HH. 1993 Molecular definition of the extreme size polymorphism in apolipoprotein(a). Hum Mol Genet. 2:933–940. 11. Boerwinkle E, Leffert CC, Lin J, Lackner C, Chiesa G, Hobbs HH. 1992 Apolipoprotein(a) gene accounts for greater than 90% of the variation in plasma lipoprotein(a) concentrations. J Clin Invest. 90:52– 60. 12. Eden S, Wiklund O, Oscarsson J, Rosen T, Bengtsson BA. 1993 Growth hormone treatment of growth hormone-deficient adults results in a marked increase in Lp(a) and HDL-cholesterol concentrations. Arterioscler Thromb. 13:296–301. 13. Tonstad S, Sundt E, Ose L, et al. 1996 The effect of growth hormone on low density lipoprotein cholesterol and lipoprotein(a) levels in familial hypercholesterolemia. Metabolism. 45:1415–1421. 14. Nolte W, Ra¨disch C, Armstrong VW, Hu¨fner M, Von zur Mu¨hlen A. 1997 The effect of recombinant human GH replacement therapy on lipoprotein(a) and other lipid parameters in adults wih acquired GH deficiency: results of a double-lind and placebo-controlled trial. Eur J Endocrinol. 137:459 – 466. 15. Querfeld U, Haffner D, Wu¨hl E, et al. 1996 Treatment with growth hormone increases lipoprotein(a) serum levels in children with chronic renal insufficiency. Eur J Pediatr. 155:913–918. 16. Laron Z, Wang XL, Klinger B, Silbergeld A, Wilcken DEL. 1997 Growth hormone increases and insulin-like growth factor-I decreases circulating lipoprotein(a). Eur J Endocrinol. 136:377–381. 17. Burman P, Johansson AG, Siegbahn A, Vessby B, Karlsson FA. 1997 Growth hormone (GH)-deficient men are more responsive to GH replacement therapy than women. J Clin Endodocrinol Metab. 82:550 –555. 18. Russel-Jones DL, Watts GF, Weissberger A, et al. 1993 The effect of growth hormone replacement on serum lipids, lipoproteins, apolipoproteins and cholesterol precursors in adult growth hormone-deficient patients. Clin Endocrinol (Oxf). 41:345–350. 19. Hassan HMS, Kohno H, Kuromaru R, Honda S, Ueda K. 1996 Body composition, atherogenic risk factors and apolipoproteins following growth hormone treatment. Acta Paediatr. 85:899 –901. 20. Webster JM, Stewart M, Al-Maskari M, et al. 1997 The effect of growth hormone replacement therapy for up to 12 months on lipoprotein(a) in growth hormone-deficient adults. Atherosclerosis. 133:115–121. 21. Tam SCF, Pang RWC, Janus ED, Kwan EYW, Low LCK. 1997 Growth hormone therapy transiently increases apolipoprotein(a) in short b-thalassaemia major children with normal growth hormone reserve. Atherosclerosis. 128:175–182. 22. Soma MR, Baetta R, Crosignani P. 1997 The menopause and lipid metabolism: strategies for cardiovascular disease prevention. Curr Opin Lipidol. 8:229 –235. 23. Henriksson P, Angelin B, Berglund L. 1992 Hormonal regulation of serum Lp(a) levels. J Clin Invest. 89:1166 –1171. 24. Stahnke N, Stubbe P, Keller E. 1992 Recombinant human growth hormone and oxandrolone in treatment of short stature in girls with Turner syndrome. Horm Res. 37:37– 46. 25. Stahnke N, Stubbe P, Keller E, Heisel HJ, Serono Study Group Hamburg. 1991 Effects and side-effects of GH plus oxandrolone in Turner syndrome. In: Ranke MB, Rosenfeld RG, eds. Turner syndrome: growth promoting therapies. Amsterdam: Excerpta Medica; 241–247. 26. Utermann G, Menzel HJ, Kraft HG, Duba HC, Kemmler HG, Seitz C. 1987 Lp(a) glycoprotein phenotypes. Inheritance and relation to Lp(a)-lipoprotein concentrations in plasma. J Clin Invest. 80:458 – 465. 27. Frison L, Pocock SJ. 1992 Repeated measures in clinical trials: analysis using mean summary statistics and its implications for design. Stat Med. 11:1685–1704. 28. Chow S-C, Liu J-P. 1998 Analysis of continuous data. In: Chow S-C, Liu J-P, eds. Design and analysis of clinical trials. Concepts and methodologies, 1st ed. New York: Wiley and Sons; 259 –305. 29. Lanes R, Gunczler P, Palacios A, Villaroel O. 1997 Serum lipids, lipoprotein lp(a), and plasminogen activator inhibitor-1 in patients with Turner’s syndrome before and during growth hormone and estrogen therapy. Fertil Steril. 68:473–477. 30. Johansson G, Oscarsson J, Rosen T, et al. 1995 Effects of 1 year of growth hormone therapy on serum lipoprotein levels in growth hormone-deficient adults. Influence of gender and apo(a) and apo(E) phenotypes. Arterioscler Thomb Vasc Biol. 15:2142–2150. 31. Mooser V, Hobbs HH. 1997 Lipoprotein(a) and growth hormone: is the puzzle solved? Eur J Endocrinol. 137:450 – 452. 32. Kronenberg F, Utermann G, Dieplinger H. 1996 Lipoprotein(a) in renal disease. Am J Kidney Dis. 27:1–25. 33. Olivecrona H, Ericsson S, Berglund L, Angelin B. 1993 Increased concentrations of serum lipoprotein(a) in response to growth hormone treatment. Br Med J. 306:1726 –1727. 34. Kronenberg F, Trenkwalder E, Dieplinger H, Utermann G. 1996 Lipoprotein(a) in stored plasma samples and the ravages of time. Why epidemiological studies might fail. Arterioscler Thromb Vasc Biol. 16:1568 –1572. 35. Evans RW, Sankey SS, Hauth BA, Sutton-Tyrrell K, Kamboh MI, Kuller LH. 1996 Effect of sample storage on quantitation of lipoprotein(a) by an enzymelinked immunosorbent assay. Lipids. 31:1197–1203. 36. Craig WY, Poulin SE, Forster NR, Neveux LM, Wald NJ, Ledue TB. 1992 Effect of sample storage on the assay of lipoprotein(a) by commercially available radial immunodiffusion and enzyme-linked immunosorbent assay kits. Clin Chem. 38:550 –553. 37. Sgoutas DS, Tuten T. 1992 Effect of freezing and thawing of serum on the immunoassay of lipoprotein(a). Clin Chem. 38:1873–1877.