点击显示 收起
1 From the Department of Maternity and Child Nursing, the School of Allied Health Sciences, Faculty of Medicine, Osaka University, Japan; and the Departments of Pediatric Surgery and Radiology, Osaka University Medical School, Japan.
2 Supported by a Grant-in-aid for Scientific Research (09470249) from the Ministry of Education. 3 Address reprint requests to Y Takagi, Department of Maternity and Child Nursing, School of Allied Health Sciences, Faculty of Medicine, Osaka University 1-7, Yamadaoka, Suita, Osaka, Japan, 565-0871. E-mail: takagi{at}sahs.med.osaka-u.ac.jp.
ABSTRACT
Background: There are no accurate indexes for determining the status of manganese in humans, and there is no clear recommended daily dose of this essential trace element to be administered in total parenteral nutrition solutions.
Objective: The objectives were to evaluate accurate indexes of manganese status and elucidate the optimal manganese dose to be administered to adult patients undergoing home parenteral nutrition.
Design: Patients were administered total parenteral nutrition solutions providing 0, 1, 2, or 20 µmol Mn/d according to an on-off design, after which manganese concentrations in whole blood and plasma were determined. Magnetic resonance imaging (MRI) was performed to determine the intensity on T1-weighted images (MRI intensity) and T1 values in the globus pallidus. Hematologic and biochemistry tests were also performed.
Results: High degrees of correlation were found between whole-blood manganese concentrations and both MRI intensity (r = 0.7728) and T1 values (r = -0.7519) in the globus pallidus. A strong negative correlation was found between MRI intensity and T1 values (r = -0.8407). The dose of 1 µmol Mn/d caused no change in MRI intensity or T1 values, and the whole-blood manganese concentration remained within the normal range in all patients.
Conclusions: Whole-blood manganese concentrations and MRI intensity and T1 values in the globus pallidus are useful indexes of the status of manganese in humans. The optimal dose of manganese may be 1 µmol/d for adult patients undergoing home parenteral nutrition.
Key Words: Home parenteral nutrition • total parenteral nutrition • magnetic resonance imaging • manganese • trace elements • nutritional requirements • optimal dose
INTRODUCTION
Manganese is considered an essential trace element for humans (1) and is ordinarily administered in total parenteral nutrition (TPN) solutions, but the kinetics of manganese in the body have yet to be elucidated. Moreover, there are no generally accepted indexes for use in determining the status (excess or deficiency) of manganese in the body. There is also no clear, standard recommended daily dose of this element. The published literature indicates a broad, 200-fold range in the recommended daily manganese dose for adults, extending from a low dose of 0.18–0.91 µmol (0.01-0.05 mg) (2) to a high dose of 40 µmol (2.2 mg) (3).
The past decade has brought reports that magnetic resonance imaging (MRI) can detect accumulation of manganese in the brain and that T1-weighted magnetic resonance images have shown high intensity in the basal ganglia, especially in the globus pallidus, of patients receiving TPN. These findings are thought to be due to excess administration of manganese (4–7). However, there have been no reports of clinical studies on relations or correlations among blood manganese concentrations, intensity on T1-weighted images (MRI intensity), and T1 values.
Accordingly, the present study was designed to elucidate MRI intensity and T1 values in the globus pallidus of patients undergoing home parenteral nutrition (HPN) and to investigate the relations between blood manganese concentrations and these 2 MRI variables at the time of TPN administration. We also evaluated the optimal intravenous dose of manganese on the basis of these variables. This is the first report of a dose-response study of manganese in HPN patients.
SUBJECTS AND METHODS
Subjects
This study was conducted between May 1994 and March 1999 in 12 patients (6 men and 6 women) undergoing long-term HPN under the care of the Department of Pediatric Surgery, Osaka University Medical School, Japan. None of the subjects had serious liver dysfunction at the time of enrollment. Patient characteristics at the time of enrollment, including information on manganese administration, are shown in Table 1. The patients ranged in age from 17 to 62 y. Four patients had short-bowel syndrome, 3 had Crohn disease, 2 had chronic idiopathic intestinal pseudoobstruction syndrome, 1 had sitophobia, 1 had enteritis, and 1 had intestinal ulcers.
View this table:
TABLE 1 . Background characteristics of subjects1
The control subjects used to establish normal blood manganese concentrations were healthy volunteers (whole blood: 25 men aged 26-46 y and 21 women aged 24-52 y; plasma: 20 men aged 28-46 y and 20 women aged 23-46 y). The control subjects used in the MRI studies of the globus pallidus had been administered an elemental diet (ED patients) daily for (1 y before entry into the present study (2 men and 3 women aged 26-60 y;
In Japan, the only commercially available trace element preparation provides the following elements daily: 20 µmol Mn, 35 µmol Fe, 5 µmol Cu, 60 µmol Zn, and 1 µmol I. For our study, the Osaka University Hospital changed the concentration of manganese while keeping the compounded amounts of the other 4 elements constant by varying the proportion of a solution containing no manganese (FeCl2, CuCl2, ZnSO4, and KI) and a solution containing only manganese (MnCl2).
In our study, 9 patients were administered manganese in a dose sequence of 20 0 2 1 µmol/d and 3 patients received manganese in a dose sequence of 20 2 1 µmol/d (Table 1). Before the study began, all 12 patients had received TPN supplemented with a trace element preparation that included 20 µmol Mn/d, and 9 of the 12 patients had undergone an on-off study in which the whole-blood and plasma manganese concentrations and MRI intensity of the globus pallidus were determined at 20 or 0 µmol Mn/d (8). The reproducibilities and reversibilities of these 3 variables and of T1 values were investigated. On the basis of Shike et al's (9) recommendation that 0.18-1.82 µmol Mn/d (0.01-0.1 mg Mn/d) be added to TPN solutions, we selected 2 µmol Mn/d as the next dose to investigate. Because the MRI intensity in the globus pallidus increased slightly after this dose, we reduced the next dose to be investigated to 1 µmol Mn/d.
Measurement of variables
The clinical variables of each patient were monitored, and whole-blood and plasma manganese concentrations were measured at 1-mo intervals. In addition, serum variables were measured (iron, copper, zinc, unbound iron binding capacity, and ferritin) and hematologic (eg, white blood cells, red blood cells, platelets, hemoglobin, hematocrit) and blood biochemistry (eg, glucose, electrolytes, blood urea nitrogen, creatinine, aspartate aminotransferase, alanine aminotransferase, -glutamyltransferase, bilirubin, C-reactive protein, total protein, albumin, lactate dehydrogenase, leucine aminopeptidase, alkaline phosphatase, cholinesterase, creatine kinase, total cholesterol, ester cholesterol, amylase, lipase, triacylglycerol, phospholipids, fatty acids) variables were measured about every 2-3 mo.
Blood was collected from all subjects with use of a disposable polypropylene syringe and a needle (Terumo, Tokyo). Whole-blood and plasma manganese concentrations were measured according to the procedures described by Matsuda et al (10, 11) with use of an atomic absorption spectrophotometer equipped with a graphite furnace (model 5700 or Z-8100; Hitachi, Tokyo) after dilution with 0.5% Triton X-100 (Wako Pure Chemical Industries, Osaka, Japan). To avoid contamination by manganese, all tools used for sampling and analysis (except for syringes and needles) were submerged in 6 mol HNO3/L for (7 d before use. To ensure quality control, bovine liver 1577b and bovine serum standard reference material (SRM-1598; National Institute of Standards and Technology, Gaithersburg, MD) were analyzed along with the whole-blood and plasma samples.
MRI of the brain was performed about every 2-3 mo with a Magnetom H15 instrument (Siemens, Erlangen, Germany) with use of repetition times (TR) of 660 and 2500 ms and an echo time (TE) of 15 ms. Radiologists rated the intensity of the images as high when the globus pallidus showed a high-intensity signal on the T1-weighted image, as moderate when the increase in intensity was slight, and none when there was no increase in intensity. The T1 value was calculated as follows:
RESULTS
Clinical findings
During the course of this study, none of the patients showed any clinical symptoms of manganese intoxication (eg, Parkinsonian-like neurologic abnormalities) or deficiency (eg, abnormalities in lipid metabolism) that could be attributed to manganese administration. In addition, none of the patients developed serious liver dysfunction [aspartate aminotransferase >100 U/L, alanine aminotransferase >100 U/L, and total bilirubin >3.4 µmol/L (2 mg/dL)]. None of the serum, hematologic, or biochemical variables was dependent on the administered manganese dose. Whole-blood and plasma manganese concentrations, MRI intensity and T1 values, and manganese doses for a representative patient in the study (ie, patient 2) are shown in Figure 1.
FIGURE 1. . Whole-blood and plasma manganese concentrations, intensity of magnetic resonance imaging (MRI) and T1–weighted magnetic resonance images in the globus pallidus, and manganese doses in a typical male patient (aged 34 y) who underwent home parenteral nutrition in July 1994. The gray-shaded area in the top 2 panels indicates our normal ranges.
HPN patients administered a trace element preparation containing 20 µmol Mn/d had a significantly higher whole-blood manganese concentration than did the ED patients and healthy control subjects (Figure 2). In contrast, the plasma manganese concentration was significantly lower in the ED patients than in the HPN patients or healthy control subjects; the difference, however, was not significant between the latter 2 groups. MRI intensity in the globus pallidus was significantly higher and the T1 value was significantly longer in the ED patients than in the HPN patients.
FIGURE 2. . Mean (±SD) whole-blood and plasma manganese concentrations, intensity of magnetic resonance imaging (MRI) in the globus pallidus, and T1-weighted magnetic resonance images in the globus pallidus of home parenteral nutrition patients (HPN) who received a trace element preparation (20 µmol Mn/d), of patients administered an elemental diet (ED) daily for 1 y before entry into the present study, and of healthy control subjects. Bars with different letters are significantly different, P < 0.05 (Tukey-Kramer multiple comparisons test). *Significantly different from ED, P < 0.05 (Wilcoxon's test). n in brackets.
Effects of manganese in HPN patients
The manganese concentration in the whole blood of HPN patients increased in a dose-dependent fashion (Figure 3). Significant differences were shown between all pairs of doses except 1 and 2 µmol Mn/d. The plasma manganese concentration (normal range: 0.0346–0.105 µmol/L, or 1.9-5.8 µg/L) of HPN patients was significantly higher at 20 µmol Mn/d than at 0, 1, or 2 µmol Mn/d. MRI intensity of the globus pallidus in HPN patients increased in a manganese dose-dependent fashion. None of the patients had an MRI intensity at the dose of 0 µmol Mn/d, 2 patients had a moderate MRI intensity at 1 µmol Mn/d, 6 patients had a moderate MRI intensity at 2 µmol Mn/d, 1 patient had a moderate MRI intensity at 20 µmol Mn/d, and 9 patients had a high MRI intensity at 20 µmol Mn/d. The T1 value became shorter in a dose-dependent fashion, and differences were significant between all dose combinations except 0 and 1 µmol Mn/d.
FIGURE 3. . Mean (±SD) whole-blood and plasma manganese concentrations, intensity of magnetic resonance imaging (MRI) in the globus pallidus, and T1-weighted magnetic resonance images in the globus pallidus of home parenteral nutrition patients who received a trace element preparation [0, 1, 2, or 20 µmol Mn/d via total parenteral nutrition (TPN)]. Bars with different letters are significantly different, P < 0.05 (Tukey-Kramer multiple comparisons test). n in brackets.
The distribution of whole-blood manganese concentrations in the HPN patients, ED patients, and healthy control subjects are shown in Figure 4. Ninety-one percent and 17% of the HPN patients had a concentration that exceeded the normal range (0.0947–0.437 µmol/L, or 5.2–24.0 µg/L) at doses of 20 and 2 µmol Mn/d, respectively. Conversely, at doses of 0 and 1 µmol Mn/d, the whole-blood manganese concentration of all HPN patients was within the normal range. However, at 0 µmol Mn/d, the whole-blood manganese concentration of all HPN patients was below the mean concentration.
FIGURE 4. . Distribution of whole-blood manganese concentrations in home parenteral nutrition patients who received a trace element preparation [0, 1, 2, or 20 µmol Mn/d via total parenteral nutrition (TPN)], of patients administered an elemental diet (ED) daily for 1 y before entry into the present study, and of healthy control subjects.
As shown in Figure 5, cases of no and high MRI intensity in the HPN patients were clearly separated on the basis of T1 values. However, the cases of moderate MRI intensity overlapped both the cases with high intensity and those with no intensity. For example, 2 HPN patients administered 1 µmol Mn/d had moderate MRI intensity and their T1 values overlapped with the distribution of T1 values recorded at the dose of 0 µmol Mn/d.
FIGURE 5. . Magnetic resonance imaging (MRI) intensity on T1-weighted images and T1 values in the globus pallidus of home parenteral nutrition patients who received a trace element preparation providing 0 (), 1 (•), 2 (), or 20 () µmol Mn/d.
Correlation coefficients
Correlation coefficients among whole-blood and plasma manganese concentrations, MRI intensity, and T1 values are shown in Table 2. Four correlation coefficients (r) were significant. MRI intensity was positively correlated with the whole-blood manganese concentration (r = 0.7728), and the T1 value was negatively correlated with the whole-blood manganese concentration (r = -0.7519), the plasma manganese concentration (r = -0.3598), and MRI intensity (r = -0.8407).
View this table:
TABLE 2 . Correlation coefficients among whole-blood and plasma manganese concentrations, magnetic resonance imaging (MRI) intensity in the globus pallidus, and T1-weighted magnetic resonance images in the globus pallidus in patients undergoing home parenteral nutrition
DISCUSSION
This is the first report of an investigation of the in vivo indexes of manganese status in the human body and of the determination of the optimal dose of manganese via changes in the administered dose of manganese in long-term TPN patients.
The signal intensity in MRI is dependent on (4 variables intrinsic to tissues: the proton density, T1 relaxation time, T2 relaxation time, and flow. T1-weighted images show contrast mainly due to differences in the T1 value. It is known that the intensity on T1-weighted images increases as the T1 value becomes shorter.
Manganese is a paramagnetic substance, and it is generally accepted that paramagnetic substances may shorten the T1 value. In an in vitro study using manganese and trace element (iron, copper, zinc, and iodine) solutions diluted with physiologic saline or rat brain homogenate, Chaki et al (12) observed concentration-dependent signal hyperintensity only for the solutions that contained manganese; no effect was observed for solution without manganese. The same team also reported that certain brain sites in rats that received TPN showed a strong positive correlation between whole-blood manganese concentrations and the signal intensity on T1-weighted images (13). Thus, the general consensus is that during parenteral nutrition, T1 shortening is due to manganese accumulation.
In the present study, the mean MRI intensity and T1 value in the globus pallidus showed a strong and negative correlation (Table 2). These 2 variables changed in a manganese dose-dependent manner when the administered doses of the other 4 trace elements (iron, copper, zinc, and iodine) were kept constant. The T1 value is influenced by both the conditions of measurement and the equipment used. Therefore, we surmise that, as long as a patient is evaluated under the same conditions of measurement and with the same equipment, the T1 value in the globus pallidus is more objective than are MRI findings as an index of the in vivo manganese status.
Whole-blood manganese concentrations in our 40 healthy volunteers was similar to other reported values (11, 14–18). On the other hand, the range of plasma manganese concentrations in these volunteers (0.0346–0.105 µmol/L, or 1.9-5.8 µg/L) was 2-3 times that of other reported values (11, 14–20). The reason for this difference is not clear; there may have been something atypical about the plasma manganese concentrations in our control group. Even so, one of our most important findings was that the whole-blood manganese concentration correlated strongly with both MRI intensity and the T1 value, whereas the plasma manganese concentration correlated weakly with both of these MRI variables (Table 2). In addition, Alves et al (19) reported that high intensity was observed even when the plasma manganese concentration was maintained in the normal range. We thus surmise that the manganese concentration in whole blood is superior to that in plasma as an index of MRI intensity and the T1 value; other investigators showed that it is also superior as an index of the manganese status in the body (14–16, 19, 21).
On the basis of these findings, we conclude that the whole-blood manganese concentration, MRI intensity in the globus pallidus, and the T1 value in the globus pallidus are useful indexes of in vivo manganese status. All 3 variables changed in a manganese dose-dependent manner. The T1 value was significantly shorter when HPN patients were administered 2 or 20 µmol Mn/d than when administered 0 or 1 µmol/d, which suggests that the manganese concentration in the brain increased at the higher doses. Fitzgerald et al (21) postulated that the optimal dose of manganese for HPN patients is 2 µmol/d. However, Bertinet et al (16) reported that when the manganese dose for HPN patients was 2 µmol/d, the T1-weighted images showed a high-intensity signal in the basal ganglia of 10 patients (71%) and that 1.82 µmol Mn/d (0.1 mg Mn/d) is probably still excessive.
The results depicted in Figure 5, the fact that the differences in T1 values at 0 and 1 µmol Mn/d were not significant, and the finding that whole-blood manganese concentrations were within the normal range in all HPN patients at 0 or 1 µmol Mn/d (although the mean concentration was significantly higher at 1 than at 0 µmol Mn/d) confirm the efficacy of manganese administered at a dose of 1 µmol/d. This dose is within the range recommended by the American Medical Association (0.18-0.91 µmol/d, or 0.01-0.05 mg/d; 2) and by the American Society for Parenteral and Enteral Nutrition (1.01-1.82 µmol/d, or 60-100 µg/d; 22).
We determined the concentration of manganese received as a contaminant in the parenteral solutions used. The manganese concentration was 0.0364–0.0728 µmol/L (2-4 µg/L) in the glucose and electrolyte solutions and was <0.0091 µmol/L (0.5 µg/L) in the amino acid solutions. The manganese concentration received by our HPN patients as contaminants (1200 mL glucose and electrolyte solution + 600 mL amino acid solution each day) was thus 0.05-0.11 µmol/d (3-6 µg/d). Isegawa et al (23) reported that the level of contamination in the solutions commercially available in Japan was <0.0275 µmol Mn/L (1.5 µg Mn/L) for lipid solutions and <0.0018 µmol Mn/vial (0.1 µg Mn/vial) for vitamin, insulin, and heparin preparations. In addition, manganese elucidated from and adsorbed to bags, infusion devices, catheters, and filters commercially available in Japan totals <0.018 µmol/L, or <1 µg/L (J Isegawa, Ajinomoto Co, Inc, unpublished observations, 2001). Thus, manganese contamination was relatively small.
Because none of the patients showed any clinical (including neurologic) symptoms suggestive of manganese intoxication or deficiency, it may be that functional changes are less sensitive to in vivo manganese status than are MRI-detected changes.
On the basis of our findings, we conclude that the whole-blood manganese concentration, the MRI intensity in the globus pallidus, and the T1 value in the globus pallidus are useful indexes for evaluating the manganese status in humans. We propose that the optimal dose of manganese in TPN solutions is 1 µmol/d for adult HPN patients.
REFERENCES