Original Article

Influence of Levels on Serum Levels

Munetada OlMOMl, Shogo MASUTA, Makoto SAKAI, Takeshi OHARA, Fumihiko HATA and Shigeaki BABA

The serum fructosamine level is influenced by the serum protein level. Therefore, we investigated the relationship between serum fructosamine levels and serum total protein and albumin levels measured simultaneously in non-diabetic and diabetic patients. Significant positive correlations were found between them. A correction formula was constructed to express the serum fructosamine level when the serum total protein level was 7 g/dl or when the level was 4 g/dl: corrected fructosamine (mmol/1) = measured fructosamine (mmol/1) + A (7 (g/dl) or 4 (g/dl) - serumtotal protein (g/dl) or serum albumin (g/dl)) where A is 0.28 for serum total protein and 0.30 for serum albumin in non-diabetic subjects; 0.29 for serum total protein and 0.35 for serum albumin in diabetic patients; and 0.29 for serum total protein and 0.31 for serum albumin in non-diabetic and diabetic subjects combined. Key words: Glycation, Glycated serum protein, Diabetes mellitus, Blood glucose control

Serum fructosamine has been clinically used as mmol/1. The serum total protein, and serum an indicator of blood glucose control in the previous albumin levels were measured in the samples which two weeks. The fructosamine level is measured by were used for the measurement of serum colorimetry (1), taking advantage of the fact that fructose-lysine, which consists of glucose bound to For studies in vitro, glucose at concentrations of a lysine residue of serum protein, is reduced under 00 mg/dl, 200 mg/dl, and 300 mg/dl was added alkalineconditionsand, therefore, the level may be to humanserumalbumin at various concentrations affected by the level of serum (2, 3). Thus, under sterile conditions. The mixture was then in- the serum fructosamine level should be expressed cubatedin 0.5 M sodium phosphate with 3 mM of after correction for serum protein level (4). Howey odiumazide, pH 7.45, at 37°C for 2 weeks. 1 fructo samine. et al. (5) have already constructed a correction s Serumsamples were used for the measurement formula for calculating the serum fructosamine of fructosamine. Fructosamine was measured level, taking the serum protein levels into considera- according to the method of Johnson et al. (1) by the tion. Wehave undertaken studies in vitro and in vivo fructosamine test (Hoffman La Roche, Basel). As to construct a correction formula for calculating the a primary standard, 1 -deoxy- l -morpholino-fructose serumfructosamine level in a similar manner. wasused. Twenty microliters of serum were colored with0.25 mmol/1 of nitro-blue tetrazolium at 37°C, MATERIALS AND METHODS at pH 10.35, and absorbance was measured at 550 The study was conducted with 175 non-diabetic nm using an automated analyzer (Cobas Mira, subjects with fasting plasma glucose ^ 100 mg/dl Japan Roche, Tokyo). and 155 diabetic patients with fasting blood glucose Serum glucose was measured by the glucose ^ 120 mg/dl and serum fructosamine levels < 4.0 oxidase method and HbAlc was measured by high-

From The Second Department of Internal Medicine, Kobe University School of Medicine, Kobe, Japan Received for publication October ll, 1988. eprint request to: Munetada Oimomi MD, The Second Department of Internal Medicine,

R Kobe University School of Medicine, 5-1, 7-chome, Kusunoki-cho, Chuo-ku, Kobe, 650, Japan

312 Jpn J Med Vol 28, No 3 (May, June 1989) Serum Fructosamine Assay performanceliquid chromatography (HPLC) using level and serum albumin level were significantly and the column chromatography system Auto Alc positively correlated with the serum fructosamine (Kyoto Daiichi Co., Ltd., Kyoto, Japan). level (Y = 0.285X + 0.850 andY = 0.304X + Serum total protein and serum albumin were 1.598, respectively, where Y = serum fructosamine measuredby the Biuret method and by the brom- level, X = serum total protein level or serum cresol-purple method (TBS-80S; Toshiba Co., albumin level (Fig. 1). Therefore, the corrected value Tokyo), respectively. of serum fructosamine can be expressed as follows: To test differences between correlation co- corrected value (mmol/1) = measured value efficients,normal approximation after z-transfor- (mmol/1) -A(7 (g/dl) or 4 (g/dl) - serum total mation ofStudent's t test was used. protein (g/dl) or albumin (g/dl)) where A is the gradient of the regression line, and 7 and 4 are the RESULT S target values of the corrected serum total protein and In non-diabetic subjects, the serum total protein Table 1. Fructosamine correction formula

4 .0 C or re c ted fructosamine = (m ea s u re d f ru ct o sa m in e + A (7.0 o r 4.0- s er um to ta l p ro te in N = 1 7 5 m m o l/i? m m o l / f i o r se ru m albu m in co n c e n tra tio n , a /d JG)) m m o l/ ｣ Y = 0 . 2 8 5 X + 0 . 8 5 0 R = 0. 6 3 1 A P < 0 . 0 0 1 3 .0 s e r u m t o t a l p r o t e in s e r u m a lb u m in CD c : ! ; ii ｣ fl N o n - d fb jeecttics o .2 8 0 .3 0 s u I/) o 3 2 .0

D ia b e t ic p 13 p a t ie n t s 0 .2 9 0 .3 5 en0) ( F r u c t o s a m i

A l l s u b j e c t s ( F r u c t o s a m i 0n. e2 <9 4 . 0 ) 0 .3 1

' 5 6 7 8 9 s e ru m t o t a l p ro t e in S /d j2 (m m o l/｣ ) G lu c o s e

4.C N = 1 7 5 Tim o l/ JO Y = 0. 3 0 4 X + 1. 5 9 8 3 .G R = 0. 6 1 9 P < 0 . 0 0 1 3 .0 . : . : l . ! i ; ! ' ! : l i: c ｣ r f i T l l - fl I/} o e o G lu c o s e 3 u CD 3 2 0 0 m g /d jO en 1. 0 L .

G lu c o s e 1 .0 1 0 0 m g /d jg 0 1 2 3 4 5 s e r u m a lb u m in 9 / d jG G lu c o s e O m g /d ｣ Fig. 1. Correlation between serum total protein and serum albumin concentrations and serum fructosamine con- 1 . 0 2 .0 3 . 0 4 .0 5 .0 centrationin non-diabetic subjects. A lb u m in (g /d jg ) Upper panel: Correlation between serum total protein and serum fructosamine concentrations Fig. 2. Fructosamine levels after incubation of various Bottom panel: Correlation between serum albumin and concentrations of with glucose at 37°C serum fructosamine concentrations for2 weeks.

Jpn J Med Vol 28, No 3 (May, June 1989) 313 Oimomi et al

m m o l/ j2 m m o l/ JZ N = 3 3 0 Y = 0 .2 94 X 4 -0 .9 37 ^ 蝣 1 Y = 0 . 2 9 1 X + 0 . 8 8 0 4 .O h 4. 0 R = 0 . 5 2 0 c P < 0 .00 1 蝣 e P < 0 .0 0 1 ra rt l i : *.:< � �� � ' : : -J H i f u -i " �� QJ (/) e 3 2 . 0 h

5T L 9 0 5 6 7 8 9 S e r u m t o ta l p r o te in g / d jg S e r u m t o ta l p r o t e in % m

m m o l/ j2 m m o l/ J2 N = 3 3 0 .o Y = 0 .354 X + 1.624R=0.484 Y=0.30 8X +1. 6 7 6 4 .0 R = 0 .4 8 9 6 P < 0 .0 0 1 c en i P < 0 . 0 0 1 o fl If) o . ｫ ..' � �' . : ..� o o 蝣 :- -3 --_. ｣ 3 .0 ｣E 3.0 蝣 ｫ � �� tｫ � .1 ��iZ ^# �%^r T.*ｻ � 3 6 (U Z3 (/)

I n 5 n o ' j i i } 0 9 /d je 1 2 3 4 5 S e r u m a lb u m in a /d jS S e r u m a lb u m in Fig. 3. Correlation between serum total protein and serum albumin concentrations and serum fructosamine con- Fig. 4. Correlation between serum total protein and serum centration in diabetic patients. albumin concentrations and serum fructosamine concentra- pper panel: Correlation between serum total protein and tion in non-diabetic and diabetic subjects. serum fructosamine concentrations pper panel: Correlation between serum total protein and ottom panel: Correlation between serum albumin and serum fructosamine concentrations U B serum fructosamine concentrations ottom panel: Correlation between serum albumin and U B serumfructosamine concentrations he corrected serum albumin, respectively (Table 1). In vitro, incubation of various concentrations of serum total protein and serum albumin levels were human serum albumin with varying concentrations significantly and positively correlated with the serum fructosamine levels (Y = 0.291X + 0.880, n=33O, oft glucose showed that the levels of fructosamine increased in proportion to the concentration of r=0.52, p<0.001 and Y = 0.308X + 1.676, albumin (Fig. 2). At given concentrations of n=330, r=0.49, p<0.001, respectively) (Fig. 4) albumin, the level of fructosamine increased with (Table 1). the concentration of glucose. In particular, the pro- duction of fructosamine was accelerated with in- DISCUSSION creasingconcentration of glucose. In determining fructosamine level by the levels In diabetic patients with serum fructosamine of serum protein, the proportion of albumin in levels < 4 mmol/1, the serum total protein and serum protein, and the life span of serum protein serum albumin levels were also significantly and may influence the measured value. Of these positively correlated with the serum fructosamine variables, serum protein levels are clinically the most level (Y = 0.294X + 0.937 andY = 0.354X + important. The present studies in vivo and in vitro 1.624, respectively) (Fig. 3) (Table 1). revealed that the serum protein level influences the In non-diabetic subjects and diabetic patients, the fructosamine level. To correct for changes in serum

314 Jpn J Med Vol 28, No 3 (May, June 1989) Serum Fructosamine Assay protein levels, a division of the fructosamine level each patient, including patients who have been sub- by the total protein level may be considered. j ected to immediate control of such as However, as seen from the relationship between the is required after diabetic ketoacidosis. In patients serum total protein or serum albumin level and the withserum fructosamine level 2^ 4 mmol/1, some serum fructosamine level, if the serum fructosamine medical treatment is immediately given to lower level is simply divided by the total protein or serum blood glucose levels. Therefore, the correction is not albumin level, the serum fructosamine level is high always necessary in patients with fructosamine level when total protein or serum albumin level is low and 2t4mmol/1. Furthermore, in patients with serum islow in the reverse case. We considered that, as fructosamine ^ 4 mmol/1, the coefficient A may Howey et al. (5) reported, a correction using fixed be large. In contrast, if serum fructosamine is less levels of total protein and albumin might be than4 mmol/1, the use of the same correction effective. As in our study in vitro, glycation might formula as in non-diabetic subjects seems to be be accelerated by evaluating the glucose level and, appropriate. therefore, we divided our subjects into non-diabetic and diabetic patients. The gradient of the linear REFERENCES regressionline for levels of protein and fructosamine was greater for diabetic patients than for non- 1) Johnson RN, Metcalf PA and Baker JR: Fructosamine: diabetic subjects. Diabetic patients were restricted a newapproach to the estimation of serum glycosyl- protein. An index of diabetic control. Clin Chim Acta to those who had serum fructosamine levels below 127: 87, 1982. 4mmol/1. Since no remarkable difference in A, the 2) Fliickiger T, Woodtli T and Berger W: Evaluation of gradient of the regression line, was found between the fructosamine test for the measurement of plasma in diabetic and non-diabetic groups, the coefficient protein glycation. Diabetologia 30: 648, 1987. A was estimated in non-diabetic subjects and 3) VanDieijen-Visser MP, Seynaeve C and Brombacher PJ: Influence of variations in albumin or total-protein diabetic patients having serum fructosamine levels concentration on serum fructosamine concentration. of < 4 mmol/1 and was found to be 0.29 for serum lin Chem 32: 1610, 1986. total protein and 0.31 for serum albumin. These 4) Mac Donald D, Pang C-P, Cockram CS, Swaminathan values were considered to be of practical clinical use. R: Fructosamine measurements in serum and plasma. Clin Chim Acta 168: 247, 1987. The correction of serum fructosamine levels is re- 5) Howey JEA, Browning MCKand Fraser CG: Assay of quired when screening for diabetes, investigating serum fructosamine that minimizes standardization and diabetic complications and comparing cases, whereas matrix problems: Use to assess components of biological it is not always necessary for follow-up studies of variation.C Clin Chem 33: 269, 1987.

Jpn J Med Vol 28, No 3 (May, June 1989) 315