A N N A L S O F C L IN IC A L A N D L A B O R A T O R Y S C IE N C E, Vol. 19, N o. 2 Copyright 1989, Institute for Clinical Science, Inc. Sensitivity of Serum Fructosamine in Short Term Glycemic Control THOMAS W. PRIOR, Ph.D., JOHN F. CHAPMAN, D r.p H. and DANIEL D. BANKSON, P h.d. Division o f Laboratory Medicine, North Carolina Memorial Hospital, 1071 Patient Support Tower, Chapel Hill, NC 27514 ABSTRACT T he serum fructosam ine concentrations m easured in 64 diab etic patients correlated (r = 0.73) w ith glycated hem oglobins (HbAlc). H ow ever, 23 percent of the diabetic patients had normal fructosamine and abnorm al H b A lc levels. In order to determ ine w h eth er or not the discrepant values w ere the result of recen t glycemic regulation by the diabetic patient, fructosam ine levels of patients suffering from diabetic ketoacidosis (along with (3-hydroxybutyrate levels) were closely monitored. It was shown that short term alterations (one to three days) in serum glucose did not significantly affect fructosamine levels. Therefore, disagreem ents betw een fructosam ine and H ba lc are m ost likely due to longer term im provem ent in glucose control by the diabetic or the result of the higher imprecision of the H ba lc assay. Introduction T he m e a su re m e n t o f th e glycated serum protein fraction, term ed fructosamine, has been proposed as a means of m onitoring glucose control in diabetic patients. 2,9 Fructosam ine levels provide a m easure of short term glycemic control (one to three weeks) in diabetic patients, unlike the glycated hemoglobin (HbAlc) which reflects glucose control over a longer period (two to three months). 1,3,11 H ow ever, th e actual interval of blood glucose control and the clinical value of fructosam ine has not b e e n well established. In th e literature, th ere are conflicting reports regarding its usefulness in m onitoring diabetics. Som e authors have found a high degree of correlation e x istin g b e tw e e n fru c to sa m in e and H ba lc and have, therefore, suggested that it be used as a reliable alternative to the m ore technically dem anding, tim e consum ing, and costly H b A lc m easurem ents. 4,6,14 Others have found poor correlations betw een the two tests and reco m m en d th a t it n o t be u se d as an a lte rn a tiv e. 7,13 Finally, Je n to rp e t al recently stated8 fructosamine should be co n sid ered a unspecific tool u n til its v a lu e in c lin ic a l p ra c tic e b e c o m e s clearer. T he p u rp o se of this stu d y was to determ ine the sensitivity of fructosamine to rapid changes in glucose co ncentration and thus b etter estimate the interval 0091-7370/89/0300-0107 $01.20 Institute for Clinical Science, Inc.
108 PRIOR, CHAPMAN, AND BANKSON of glucose control that is reflected by a fructosam ine m easurem ent. An initial evaluation of the Roche fructosamine kit was also undertaken in order to assess the kit s analytical reliability. M aterials and M ethods F r u c t o s a m in e Fructosam ine was m easured spectrophotom etrically using the Roche fructosam ine kit and th e Roche Cobas BIO Centrifugal Analyzer.* The m ethod is based on the ability of fructosamines to reduce nitroblue tetrazolium (NBT) at an alkaline ph. The kit m anufacturer s protocols w ere used and the assay was calibrated using the 1-deoxy-l-morpholino- D -fru c to se s ta n d a rd (3.2 m m ol/l ) supplied by Roche. G l y c a t e d H e m o g l o b in T he concentration of glycated hem o g lo b in was d e te rm in e d by a catio n exchange chrom atography m ethod, f P - H y d r o x y b u t y r a t e (3-hydroxybutyrate (0HBA) was quantitated enzymatically with (3-hydroxybutyrate dehydrogenase using the Cobas BIO Centrifugal Analyzer. The reaction involves m easuring the increasing absorbance at 340 nm associated w ith th e reduction of nicotinam ide adenine dinucleotide. 5' 12 G l u c o s e Glucose was m easured by an oxygen electrode using glucose oxidase on an Astra 8 analyzer. $ * Roche Analytical Instruments Inc., Nutley, NJ. t Helena Laboratories, Beaumont, TX. t Beckman Instruments, Brea, CA. St a t is t ic a l M e t h o d s L inear correlations w ere estim ated using the m ethod of least squares. The u n p aired stu d e n ts t te st was used to d e te r m in e s ta tis tic a l s ig n ific a n c e betw een groups. Results and Discussion M e t h o d E v a l u a t io n R e fe r e n c e R a n g e. F ru c to s a m in e values for 40 non-diabetic individuals are shown in figure 1. The observed range was 2.14 to 2.75 m m ol p e r L, w ith a mean of 2.40 mmol per L and a standard deviation of 0.15. Although the distribution of fructosamine values is not gaussian (d e m o n s tra tin g slig h t p o sitiv e skew), nonparam etric testing yielded a similar range of 2.17 to 2.71 mmol per L. Precision. Two pools of human serum w ere used to assess within-run precision. T he CVs w ere 1.0 p e rc e n t (n = 22, m ean fructosamine = 2.44 mmol per L) and 0.89 percen t (n = 22, m ean fructo sam in e = 4.19 m m ol p e r L). The betw een-run precision was m easured using two commercially available lyophilized quality control sera. The CVs were 2.3 p e rc e n t (n = 12, m ean fru cto s am ine = 2.64 m m ol p e r L), and 1.1 p e r c e n t (n = 1 2, m e a n f r u c t o s amine = 4.15 mmol p er L). Linearity. T he linearity was d e te r m ined by serially diluting a com m ercially quality control serum with isotonic saline. The upper lim it of linearity was 5.5 mmol per L (r = 0.999). C l in ic a l C o r r e l a t io n Fructosamine Levels in Diabetics. The mean serum fructosamine concentration m easured in 64 diabetics was 3.07 mmol p e r L (SD = 0.79), w hich was significantly different from that found in the norm al group (p < 0.001). The fructos-
SENSITIVITY OF SERUM FRUCTOSAMINE 109 FRUCTOSAMINE LEVELS (mmol/l) am ine concentrations correlated signifi cantly w ith H b A lc values (r = 0.73, fig u re 2). H ow ever, w hen th e refe re n c e ranges w e re co n sid e red, discrepancy b e tw e en th e two tests was observed. Tw enty th re e p e rc e n t of th e diab etic population had elevated H b A lc values (> 6.6 percent) and norm al fructosam ine levels. It has b e e n shown that fructos am ine is d e p re sse d by low concentra tions of p ro te in.10,13 This, however, was not a factor because these patients had normal (6.8 to 8.3 g p er dl) total protein levels. Some of the disagreem ents may be analytical in nature, as a result of the greater im precision of the H ba lc m ea surem ents (CV = 9. 6 percent). The dis agreem ents may also be due to the fact that the tests m easure different periods of glucose control. Since the fructosa m ine levels are indicators of shorter term glucose control, it is possible th at the high frequency of norm al fructosam ine values w ere due to the efforts of the dia b e tic p a tie n t to gain glycem ic control shortly before the clinical visit.8 If this is th e case, th en th e serum fructosam ine may be too sensitive to recent changes in glucose levels and thus provide the phy sic ia n w ith m is le a d in g in fo rm a tio n regarding overall glycem ic control for the previous two to three weeks. F i g u r e 2. R e la tio n betw een fructosamine and glycated hemoglobin in 64 diabetic patients. The dot te d lin e s in d ic a te th e u p p er lim it of th e refer ence ranges.
110 PRIOR, CHAPMAN, AND BANKSON In order to determ ine the sensitivity of fructosam ine to changes in glucose levels, the fructosamine concentrations were studied in three diabetic patients w ho e n te re d th e hospital in d iabetic ketoacidosis and whose glucose and (3- h y d ro x y b u ty ra te w e re m o n ito re d. Ketoacidosis was clearly evident by the severely elev ated 3-hydroxybutyrate levels in all three patients (normal = 40 to 480 (xmol per L). Patient A (figure 3) had an initial glucose of 1,295 mg per dl (71.9 mmol per L), a (3-hydroxybutyrate of 13,610 (jumol p e r L, and a fru cto s amine of 5.6 mmol per L. After insulin treatm ent, his glucose had fallen to 2 1 1 m g p e r dl (1 1.6 m m ol p e r L) and rem ained less than 150 mg per dl (8.3 T i m s ( h ) F I G U R E 3. G lucose, fru c to s a m in e, and (3- hydroxybutyrate in patient A.
SENSITIVITY OF SERUM FRUCTOSAMINE 111 mmol per L) during the next 40 hours. Although there was an initial 85 percent decrease in glucose over th e first 15 hours, the fructosam ine decreased by only 25 percent (5.6 to 4.2 mmol per L) and rem ained above the reference limit throughout the next 40 hours. Patient B (figure 4) had an initial glucose of 995 mg per dl (55.2 mmol per L), (3-hydroxybutyrate of 8,306 xmol per L, and a fructosamine of 6.62 mmol per L. The glucose values flu ctu ated throughout the four day period: 153 mg per dl (8.4 mmol per L) at 13 hours, 450 mg per dl (25.0 mmol per L) at 26 hours, and 82 mg per dl (4.5 m m ol p e r L) at 70 hours. P atient B s fructosam ine initially decreased from 6.62 to 5.10 during the first 13 hours and then rem ained fairly constant during the following four days, despite the fluctuat F ig u r e 4. G lu c o s e, f r u c t o s a m i n e, a n d ß - hydroxybutyrate in p a tie n t B.
112 PRIOR, CHAPMAN, AND BANKSON F ig u r e 5. G lu c o s e, f r u c t o s a m i n e, a n d ß - hydroxybutyrate in p a tie n t C. ing glucose levels. Patient C (figure 5) entered the hospital w ith a glucose of 695 mg per dl (38.6 mmol per L) a ß- hydroxybutyrate of 11,618 xmol per L, and a fructosamine of 6.21 mmol per L. After four hours, the glucose had fallen to 185 mg per dl (10.3 mmol per L) and g radually re tu rn e d to norm al by 2 2 hours. T he fructosam ine values parallele d, to a m uch le sse r d e g re e, th e decreasing glucose levels and rem ained severely elevated after 22 hours (> 3.7 mmol per L). In conclusion, a proportion of the diabetic population (23 percent) exhibited n o rm a l fru c to sa m in e and e le v a te d H ba lc levels. Some of the disagreement may be due to the higher imprecision of
SENSITIVITY OF SERUM FRUCTOSAMINE 113 th e H b A lc test. H ow ever, ab so lu te agreem ent should not be expected, since th e tests are expressions of different asp ects of m etabolic co n tro l. I t was shown that the differences w ere most likely not th e consequence of abrupt im provem ent in glucose control before the clinic visit. Large glucose and (3- hydroxybutyrate shifts, observed in the th re e p atien ts, resu lte d in relatively m odest changes in fructosamine. Fructosam ine has been shown to be m ore responsive than H ba lc for detecting early changes in glycemic control and, therefore, can be used to assess the adequacy of glucose control over a shorter period of tim e. H ow ever, short term (one to three days) im provem ent in glucose control could not appreciably be dem onstrated by the fructosamine levels and poor glycemic regulation would not be apparent over the relatively short term. O ur experience w ith the Roche fructosam ine kit dem onstrated a procedure which was analytically reliable, correla te d reasonably w ell w ith H b A lc results on the same patients (figure 2 ), and was economical and convenient to perform. References 1. Ar m b r u ster, D. A.: Fructosamine: Structure analysis, and clinical usefulness. Clin. Chem. 33:2153-2163, 1987. 2. B a k e r, J. R., M e t c a l f, P. A., H o ld a w a y, I. M., a n d J o h n s o n, R. N.: S e ru m f ru c to s a m in e co n c e n tra tio n as a m e a su re o f blood g lu cose control in ty p e I (insulin d ep en d en t) diab etes m ellitus. B rit. M ed. J. 290:352 355, 1985. 3. Ba k er, J. R., O C o n n o r, J. P., M e t c a l f, P. A., L a w so n, M. R., and J o h n so n, R. N.: Clinical usefulness of estimation of serum fructosamine concentration as a screening test for diabetes mellitus. Brit. Med. J. 287:863 867, 1983. 4. B a k e r, J. R., R e i d, I., and H o ld a w a y, I.: Serum fructosamine in patients with diabetes m elitus. New Zealand Med. J. 98:532-535, 1985. 5. Bankson, D. D. and C hapm an, J. F.: M anuscript in preparation, 1988. 6. H in d l e, E. J., R o st r o n, G. M., C la rk, S. A., and G att, J. A.: S erum fructosam ine an d glyc a te d h a e m o g lo b in m e a s u re m e n ts in d ia b e tic c o n tro l. A rc h. D is. C h ild h o o d 6 1 :1 1 3-1 1 7, 1986. 7. H o w e y, J. E. A., B r o w n in g, M. C., and F raser, C. G.: Assay of serum fructosamine that minimizes standardization and matrix problems: Use to assess components of biological variation. Clin. Chem. 33:269-272, 1987. 8. J e r t o r p, P., S u n d k v ist, G., F e x, G., and J e p p s- SON, J-O. Clinical utility of serum fructosamine in diabetes mellitus compared with hemoglobin Ale. Clin. Chim. Acta 175:135-42, 1988. 9. J o h n s o n, R. N., M e t c a l f, P. A., and Baker, J. A.: Fructosamine: A new approach to the estimation of serum glycosylprotein. An index of diabetic control. Clin. Chim. Acta 127:87 95, 1982. 10. L e m o n, M. a n d F o r r e st, A. R. W.: F ru c to s am in e activity o f p ro te in s in se ru m C lin. C h e m. 33:2101, 1987. 11. L im, Y. S. and Staley, M. J.: Measurement of plasma fructosamine evaluated for monitoring diabetes. Clin. Chem. 31:731 733, 1985. 12. M c M urbay, C. H., B r a n c h fl o w e r, W. J., and R ic e, D. A.: A utom ated kinetic m ethod for D-3hydroxybutyrate in plasma or serum. Clin. Chem. 30:421-425, 1984. 13. M o s c a, A., C a r e n in i, A., Z o p f i, F., C a r fine l l i, A., Ba n f i, G., C e r io t t i, F., Bo n in i, P., and PozzA, G.: Plasma protein glycation as measured by fructosam ine assay. C lin. C hem. 33:1141-1146, 1987. 14. Sa n-g il, F., Sc h ie r, G. M., M o s e s, R. G., and G a n, I. E. T.: Improved estimation of fructosamine, as a measure of glycated serum protein, with the Technicon RA-1000 analyzer. Clin. Chem. 31:731-733, 1985.