Journal of Analytical Toxicology,Vol. 23, November/December1999
Determination of Germanium in Human Specimens:
AtsukoShinohara, Momoko Chiba, and Yutaka Inaba
Departmentof Epidemiologyand EnvironmentalHealth,Juntendo UniversitySchoolof Medicine, Tokyo 173-842l, Japan
I Abstract I 6 deaths, have been reported from 1982 to 1988 (3). These hap-
penings were associatedwith Ge-containingdietarysupplements.
The determination methods of germanium (Ge) in biological Ge was believedto have health benefits because some medicinal
specimens such as blood plasma, erythrocytes, urine, hair, nail, and plants have high Ge content and antitumor or immunomodula-
other organs were established using graphite furnace atomic tion properties of several organogermanium compounds have
absorption spectrometry (GFAAS) and microwave-induced plasma been reported (4,5). Ge-containingtablets or elixirshave not been
mass spectrometry (MIP-MS). The detection limits of Ge standard
sold in general drug stores in Japan since an administrativeguid-
solution were 3 ng/mL with GFAAS and 0.05 ng/mL with MIP-MS.
The detection limits in organ samples depended on the type of
ance in 1988. However,several cases of Ge-inducedrenal toxicity,
samples and sampling amounts: 3-30 ng/g by GFAASand neurotoxicity,and renal failure of diabetics have been reported in
0.05-0.5 ng/g by MIP-MS. The sensitivity of GFAAS was lower than subsequent years (6,7). Most of the Ge intoxication was induced
that of MIP-MS; however, it was adequate for determining Ge by the intake of inorganic germanium oxide (GeO2). Renal dys-
concentrations in specimens from patients who had ingested Ge. function was also induced with organogermanium compounds
Samples were digested by a simple wet-ashing procedure using nitric such as germanium lactate citrate (8). Until recently, another
acid and perchloric acid. To avoid the interfering effects of coexisting organic Ge compound, carboxyethylgermaniumsesquioxide(Ge-
elements and perchloric acid residue, an extraction method using 132), was used as a drug for research and was administered to
organic solvent was tried. When using MIP-MS, extraction was not patients in Japan suffering from cancer or rheumatism. In addi-
necessary; however, both dilution and addition of an internal standard tion, 3-oxygermylpropionicacid polymer (propagermanium)has
were needed. Special attention was required for iron-rich samples
been used in Japan since 1994 for the treatment of chronic hep-
because a molecular ion of S6Fe160was observed at m/z 72 where
Z2Ge was monitored. The results of Ge concentrations in human
atitis B because of its modulating action on host responses (9,10).
samples obtained by these methods agreed well. Interfering effects of A sensitive and selective method for Ge determination in
perchloric acid, which was used for digestion and which remained in human specimens is necessaryfor checking the accumulation of
samples, were observed in both methods. Hair and nail samples from Ge in the body to avoid intoxication and for checking adverse
people who had ingested Ge were useful for monitoring Ge in the effects of Ge-containing compounds or drugs. Organ samples
body. Hair samples were useful for determining past exposure to Ge obtained by biopsyor by autopsyprovidedirect information about
when the distribution patterns from the scalp to the end of the strand Ge concentrations; however,these samples are available only in
were analyzed. In control subjects, Ge concentrations in the listed special cases. Bloodand urine are common samples for diagnostic
specimens and organs were lower than 0.1 pg/g or mL, and these low purposes and are considered to be good indicators reflecting the
levels of Ge were able to be determined by MIP-MS in combination present state of Ge in the body. Hair and nail samples can be
with the extraction method. obtained painlessly and are expected to indicate recent or past
exposure to Ge. These biological specimens contain alkali and
alkaline earth elements at high concentrations, and the amount
of sample availablefor analysis is often limited.
Various techniques of spectrophotometry, atomic absorption
Introduction spectrometry (AAS),and other methods have been reported for
Ge determination (11-15). High-frequency plasma mass spec-
Germanium (Ge) is used mainly in industrial fieldsand is found trometry, such as inductivelycoupled plasma mass spectrometry
ubiquitously in soil, water, plants, foods, and biomaterials (1). (ICP-MS) and microwave-induced plasma mass spectrometry
Although the toxicity of Ge is low compared with other heavy (MIP-MS) can allow sensitive and multielement analyses.
metals, there havebeen more than 30 reported human casesof Ge Graphite furnace atomic absorptionspectrometry (GFAAS)is also
intoxication (2). In Japan, 23 cases of Ge intoxication, including available and has the advantage of being widely used and
Reproduction(photocopying) editorialcontentof thisjournalis prohibitedwithoutpublisher'spermission. 625
Journal of Analytical Toxicology,Vol. 23, November/December1999
common. ICP-MSand MIP-MSequipment is relativelyexpensive 1987. Specimens from Patient A were taken during the autopsy
but has recently become available in many laboratories.With this and obtained through a physician 14 months after the patient's
in mind, we developed suitable analytical conditions for deter- death; specimens were stored at -80~ Patient B, a 60-year-old
mining Ge in biologicalmaterials using two methods: GFAASand male, was aliveand had been diagnosedwith diabetes mellitus. He
MIP-MS. These methods were applied to human specimens from had a past history of GeO2 intake over approximately 10 years.
patients who had ingested Ge compounds and to a control group. Patient C, a 48-year-oldfemale,was aliveand had ingested Ge-132
for several months in the past. Patient D, a 68-year-old male, had
taken organogermanium to maintain a healthy body and had no
symptoms of Ge toxicity.In order to obtainsamples from patients,
Experimental informed consent and approval according to the criteria of an
intramural ethical committee were obtained by individualphysi-
GFAAS (Polarized-Zeeman atomic absorption spectrometer The reagents, AAS-special-gradeHNO3,HCIO4,and HC1,analyt-
Z-6100, Hitachi, Tokyo, Japan) and MIP-MS (P-7000, Hitachi) ical-grade CCI4,and Ge standard solution (1000 IJg/mL)for AAS
were used. The operating conditions are listed in Table I. were purchased from Wako Pure Chemical Industries (Osaka,
Acetylene-flameAAS (Polarized-Zeemanatomic absorption spec- Japan). Deionized water through Milli-XQ (Millipore, Tokyo,
trometer Z-6100, Hitachi) was used for determining Na, K, Ca, Japan) with relative resistance of 18.3 M~ was used.
and Fe concentrations.
Materials Ten to 100 mg of hair, nail, and organ sample, 200-400 pL of
Nail, hair, blood, urine, and organ specimens from controls and plasma and erythrocytes, and 1 mL of urine sample were weighed
patients were analyzed. Controls had no history of ingesting Ge precisely in demetalolyzed Pyrex glass tubes. Hair and nail
compounds. Nail, hair, blood, and urine were obtained from two samples were washed with deionized water to remove external
laboratory volunteers, and nail, hair, and organ samples were contamination and were air driedbefore weighing. These samples
obtained from 12 bodies donated for medical anatomy dissection. were heated at 120-140~ with 2 mL HN03and 0.2 mL HCIO4on
The other specimens were from four patients who had ingested a block bath (Advantec,Tokyo,Japan) to digest organic compo-
Ge compounds in the past. Patient A, a 58-year-oldmale, ingested nents, then diluted with 0.5% HNO3. A series of standard solu-
GeO2 for 1-2 years and died from renal failure and myopathy in tions was prepared from the Gestandard solution by dilution with
Table I. Operating Conditions of GFAAS and MIP-MS
for Ge Determination
Resonancesource: Hollow-cathode lamp
Lamp current: 12.5 mA GFAAS
Analytical wavelength: 265.2 nm Four types of standard series as shown in Figure 1 were tried.
Background correction: Polarized-Zeeman
The calibration line of the Ge standard solution prepared with
Slitwidth: 0.4 nm
0.5% HNO3 was linear from 10 ng/mL to 1 IJg/mL when the
Graphite furnaceoperation sample volume injected into the furnace was fixed at 10 ]JL
Atomization tube: Untreatedgraphitetube (Figure 1A, line a). Ge standard solution showed a straight line
Injection volume: 10 pL over the range up to 2 !Jg/mL.Asa matrix modifier,addition of Ni
Carrier gas: Ar was tried. Addition of Ni(NO3)2solution increased the sensitivity
Heating program Temp(~ Time (s) Gas flow (mt/min) dose dependently.The optimumamounts of Ni addition were 1 IJg
Dry 80~120 30 200 or more. In the present study, 10 pL of 100 ]ag/mLNi solution was
Char 700 30 200 injected into the furnace together with 10 ]JL of the sample solu-
Atomize 2700 10 0 tion. The slope of the standard curve was about three times higher
Clean 2900 4 2O0 with Ni modifier than that obtained without modifier. Detection
limit of Ge under the present conditions was 3 ng/mL, which
corresponded to 0.03 ng.
Power: 1.3 kW A series of standard solutions were digested under the same
Plasmagas flow: 13 L/min conditions as sample digestion. When 1 mL of standard solution
Carrier gas flow: 1.3 L/min was heated with 2 mL HNO3 and 0.2 mL HCIO4at 120-140~ for
Sampling cone: Pt-made 2-3 h and the remaining residue was diluted to 1 mL with 0.5%
HNO3, the calibration line was linear, but the slope was much
cooling chamber: 4oc
lower than that without digestion (Figure 1A, line b). A loss of Ge
Scanning conditions: Dwell time:50 ms
Numberof sweeps:3 during the digestion process was observed when the remaining
Sampleflow:0.24 mL/min acid was volatized to dryness at 200~ The lost percentage of Ge
varied from 25 to 75% depending on the tube position on the
Journal of Analytical Toxicology, Vol. 23, November/December 1999
heating block. In the following experiments, temperature was 10% under the present conditions. Addition of HCIO4to the cali-
kept at 140~ for 3 h in order to have HC104remain in the tube, bration standard solutions decreased ion counts of Ge at m/z 72
and the remaining residue diluted with 0.5% HNO3was applied to depending on the amounts added, and ion counts reached 75%
GFAAS. It was found that an addition of HCIO4to Ge standard when HC104volume was 10%. This suggests that the decrease in
without heating process showed a dose-dependent decrease of ion counts of digested standards was due to the remaining HCIO4.
absorption, suggesting the interference from the remaining When the standard solutions were digested with biological sam-
HCIO4. When a series of calibration standard solutions were ples, ion counts were decreased depending on the types and
digested with 0.5 mL of urine or with 50 mg of hair from a con- amounts of samples. With 10 pL or 10 mg of plasma, erythro-
trol, calibration curveswere flat (Figure ]A, lines c and d). The Ge cytes, urine, hair, or nail sample as the original amount before
absorption was also lowered when the Ge standard solutions were digestion, the standard curves were linear, and their slopes were
mixed with digested urine or hair samples. These results indi- from 93.1 to 102.9% of digested standard without biologicalspec-
cated the interfering effects of coexisting elements in these sam- imens. In the case of plasma and urine samples, the ion counts of
ples. To avoid interfering effects of matrix elements and/or Ge at rn/z 72 decreased with increased amounts indicating the
remaining perchloric acid, an extraction method was applied. interfering effects of matrix elements (Figure 2A). In the case of
Digested standard solution with or without biological materials erythrocytes, the ion counts decreased until the addition of 50 pL;
was extracted in 1 mL CC14under 8N HCI and then re-extracted however,they increased again with the addition of 100 pL (Figure
in 1 mL deionized water according to the methods reported by
Mino et al. (12)with slight modifications.The absorbance showed Table II. Ge Concentrations in Human Samples
a linear relationship with Ge concentration (Figure 1B) when the Determined by GFAAS and MIP-MS
added amounts of urine, hair, nail, plasma, or erythrocytes were Ge (pg/mLor g)*
1.0 mL, 50 rag, 20 rag, 0.1 mL, or 0.1 mL, respectively,that is, the
original amounts beforedigestion. The slope of the digested stan- Sample GFAAS MIP-MSt MIP-MS*
dard was the same as that of the standard without digestion. The
slopes of digestedstandards with biologicalspecimenswere in the PatientA
Pancreas 2.4 • 0.9 2.78 • 0.38 2.81 • 0.51
range of 98 to 114% of the digested standards without biological
M. iliopsoas 2.4 • 0.7 2.54 • 0.75 2.50 • 0.87
specimens. Ge contents in samples were calculated using a
M. quadriceps femoris 5.1 • 1.1 4.69 • 0.43 4.62 • 0.49
matrix-matched calibration curve that contained corresponding M. pectoralis 4.7 • 1.2 4.54 • 1.20 4.96 • 1.19
amounts of the specimen obtained from controls.
Removalefficiencyof major elements in these specimens using PatientB
Serum < 0.05 0.019 • 0.003 0.018 • 0.002
this extraction method was checked by flame AAS.When 1.0 mL
Urine 0.25 • 0.06 0.226 • 0.012 0.235 • 0.008
of urine was digested, Na and K concentrations in the aqueous
Hand nail 12.6+1.9 12.97• 12.74•
phase beforeextraction were 5312 and 752 pg/mL, and those after Foot nail 16.9• 16.62+7.59 15.86+8.02
extraction were 24 and 3 pg/mL, respectively, indicating that Hair 53.1 • 5.6 56.72 • 13.43 59.62 • 13.75
more than 99.5% of Na and K were removed. When 0.1 mL of
* Mean • SD (n= 3).
plasma or erythmcyte was digested, the concentration of Na in + With dilution only.
the plasma sample and the concentrations of K and Fe in the ery- * With dilution and extraction.
throcyte sample in aqueous phase were decreased from 278, 243,
and 98 pg/mL to 1.5, 0.3, and 0.1 pg/mL, respectively,by extrac-
Table III. Ge Concentrations in Specimens from Patients
tion. The removal efficiency was more than 99.4%. Similarly, Detemined by MIP-MS with Dilution
more than 98.5% of Ca was able to be removed from nail and hair
samples by this extraction method. Ge (pg/mLor g)*
Sample PatientA PatientC PatientD
Ge has fivestable isotopes ofm/z 70, 72, 73, 74, and 76 with nat- Plasma - 0.001 • 0.001
ural abundances of 21.2, 27.7, 7.7, 35.9, and 7.4%, respectively. Erythrocytes - 0.021 • 0.007
Calibration standard curves showed linearity in the range of 0.5 to
500 ng/mL. Within these isotopes, m/z 70 coincides with one of Hair 71.05 _+6.31 4.89 + 0.53 10.96 + 0.91
the isotopes of zinc (Zn), and rn/z 74 and 76 coincide with iso- Nail (hand) 16.90 • 4.88 0.02 t 55.14 + 9.40
Nail (foot) 0.200 • 0.069 53.48 • 16.67
topes of selenium (Se). Therefore, ion counts at m/z 72 and 73
were preferentiallyused for Ge determination for Zn- and Se-con-
Spleen 49.74 + 2.93 -
taining biologicalsamples. The detection limits of Ge were 0.05
Kidney 6.15 • 1.08 -
and 0.2 ng/mL at m/z 72 and 73, respectively,when 3 a definition Lung 3.93 • 1.66 -
was used. Stomach 3.90 • 0.65 -
The effects of digestion and coexisting elements in the sample Heart 3.28 • 0.85 -
on Ge determination were also checked. Digestedstandard solu- Muscle intercostal 1.98 • 1.27
tions without biologicalsamples showed a straight line; however,
* Mean • SD (n=3).
the slope was 74.5% of that without digestion. The volume of * Mean (n=2).
HCIO4 in digested standard solutions was estimated to be about
Journal of Analytical Toxicology, Vol. 23, November/December 1999
2A). A correction based on an internal standard was performed inal amounts.
using gallium (Ga) with 10 ng/mL as the final concentration In the case of erythrocytes, the ion counts at rrgz 72 increased
because the concentrations of Ga in these biologicalsamples were with sample volume, but those at m/z 74 did not, suggesting the
under the detection limit and the mass-to-charge ratios of Ga, 69 existence of molecular ions at m/z 72. The spectra of erythrocytes
and 71, were close to the mass-to-charge ratios of Ge. Biological and other samples of a control subject were measured from ra/z
samples contained a lot of chloride, and the molecular ion 71 to 76. As shown in Figure 3, the erythrocyte showed a peak at
3~Cl1602with ra/z 69 was easily formed; therefore, the ion counts rrUz72 but did not indicate any clear peak at ra/z 74 (Figure 3, line
of ~lGa were used. The slopes of calibration lines obtained with b). With other specimens such as plasma, urine and hair, a similar
standard solutions digested with and without biological samples peak was not observed (Figure 3, lines c-e). Erythrocytes contain
were in the range of 98.5 to 101.8% of the original standard solu- much higher concentrations of iron than the other specimens.
tion. The matrix effects on the ion counts were negated by The spectra of 100-pg/mL Fe solution indicated a clear peak at
internal standard correction when the added amounts of biolog- m/z 72. The solutions of other biologicallymajor elements, Na, K,
ical samples were 50 pL or ng or less (Figure 2B). Ca, Zn, and Mg, did not indicate a matching peak, suggesting that
Matrix effects become a potential problem when samples that the peak at m/z 72 was associatedwith iron (Figure 3).
contain very low levelsof Ge must be large in measuring solution. These results indicate that the internal standard correction is
The extraction method mentioned in the GFAASsection was necessary for all samples and that the extraction method is useful
examined to decrease the effect of coexisting elements. The for samples that contained low levels of Ge. The rrgz 72 is prefer-
recovery rates of 50 ng/mL Ge from standard solutions prepared entially used for Ge determination, but rrUz 73 is recommended
with 0.5% HNO3 to final aqueous phase were 90.8 • 1.5%. The when iron-rich samples are measured without extraction.
extraction efficienciesof standard solutions digested with biolog-
ical specimens varied between 93 and 103% of standard solution Determination of Ge in patients and control subjects
only when 1 mL measuring solution contained ~ 1 mL of plasma Ge concentrations in specimens from Patients A and B were
or erythrocytes, ~ 2.5 mL of urine, or - 250 mg hair as their orig- determined by both GFAASand MIP-MS.The values obtained by
OAI A B
j 0.4 b b ~
a e w
d ~ 0.t
2oo 4oo 6oo
Ge (ng/mL) Ge (ng/mL)
Figure1. Calibration lines of Ge measured by GFAASwithout extraction (A) and with extraction (B). Identification: (a, @) standardsolution; (b, &) standard solu-
tion subjected to the same procedure for the digestion of biological samples;and standard solution digested with (c, II) 50 mg hair, (d, [-]) 1 mL urine, (e, 9 0.1
mL plasma, (f,/k) 0.1 mL erythrocytes, and (g,V) 20 mg nail.
2060o A 2.5
1,5 9 plnma
"6 1o000 > erythrocyte8
O 1 e nail
0 i i i i i 0 i L L L I
o 20 40 60 80 10o 120 0 20 40 60 80 11111 120
Amounts of biological sample added Amounts of biological sample added
(mg or pL/mL) (mg or pL/mL)
Figure2. Effectof added amountsof sampleson ion counts at m/z 72 measuredby MIP-MS (A) and thosewith internal standardcorrection (B).The concentrations
of Ge standard and Ga (internal standard)were fixed at 50 and 10 ng/mL, respectively.
Journal of Analytical Toxicology, Vol. 23, November/December 1999
GFAAS, by MIP-MS with only dilution, and by MIP-MS with symptoms of Ge intoxication; however, exposure to Ge was
extraction agreed well, as shown in Table II. Ge in serum from thought to be remarkable because of evidence of high Ge concen-
Patient B was under the detection limit by GFAAS but was trations in hair and nail (Table III).
detectable by MIP-MS.These results indicate that both methods Ge concentrations in specimens from control subjects who had
are useful for Ge determination in biological specimens and that no history of specific Ge ingestion were very low; therefore, the
the sensitivity is higher in MIP-MS than GFAAS.Ge concentra- samples were measured by MIP-MSwith extraction. As shown in
tions in other specimens from Patient A and in specimens from Table IV, Ge concentrations were lower than 4 ng/mL (or g) in
Patients C and D were measured by MIP-MSwith dilution only hair, nail, plasma, and urine from two controls. Ge concentra-
after digestion (Table III). tions in nail, kidney, spleen, liver, lung, and bone from autopsy
The Ge content in hair was the highest in the samples from control samples were lower than 70 ng/g. The averageconcentra-
Patient A,who ingested GeO2and died from myopathy and acute tions of Ge in autopsy samples were as follows: lung > kidney >
renal failure (16), as shown in Tables II and III. Among soft tis- liver, bone, spleen > nail. Gender differenceswere not clear, but
sues, the highest concentration of Ge was found in spleen, fol- average Ge concentrations in kidney were higher in females and
lowed in decending order by kidney, lung, stomach, heart, and those in bone were lower in females.
pancreas. Ge concentrations in muscles were within the same
order, where concentrations in the quadriceps femoris and pec-
toralis muscleswere about 5 IJg/g,and those in the iliopsoasand
intercostal muscles were about 2 IJg/g. Patient B had ingested Discussion
GeO2 for 10 years before stopping. Hair, nail, serum, and urine
samples were obtained 3 times at 2-3-month intervals. Table II The results of this work indicate that both GFAASand MIP-MS
shows the results of the first sampling. The concentrations of Ge were useful for Ge determination in biological specimens. The
in serum from Patient B were 0.019 + 0.003 and 0.010 • 0.005 detection limits of calibration standard solution were 3 ng/mL
I~g/mLat the second and third samplings, respectively. Ge con- with GFAASand 0.05 ng/mL with MIP-MS.The detection limits
centrations in urine were 0.039 and 0.051 IJg/mL,respectively,at in organ samples depended on the types of specimens and sam-
the second and third samplings. Ge concentrations in nail and pling amounts; when 1.0 mL of plasma, erythrocytes, or urine;
hair were not clearly changed during this period. 0.25 g of hair; or 0.1 g of nail or other organ was digested and
Patient C had ingested Ge-132 but stopped the ingestion several back-extracted in I mL deionized water, detection limits in sam-
months before the collection of blood, nail, and hair. As shown in ples were between 3 and 30 ng/g by GFAASand were between 0.05
Table III, Ge concentrations in plasma and in nails were very low. and 0.5 ng/g by MIP-MS.In MIP-MSwithout extraction, digested
The concentration of Ge in hair sample of about 10 cm in length must be diluted at least 20 times from the original weight of sam-
from the scalp from Patient C was 4.89 + 0.53 lJg/g. This hair ples and internal standard correction must be adopted to avoid
sample was cut into 1-cm pieces, and Ge concentration in each interfering effects of coexisting elements and remaining HCIO4.
piece was determined. Ge concentrations were different depend- The detection limits of MIP-MSwithout extraction were 1 ng/g.
ing on the piece, being higher at the end of the strand, then The effectsof coexistingelements and HCIO4were thought to be
decreasing to levels under the detection limit at distances of less matrix interference, decreased sensitivitybecause of the sample's
than about 3 cm from the scalp (Figure 4). Patient D had no having a greater viscosity than the standard solutions prepared
Table IV. Ge Concentrations in Specimens from Volunteers and from Bodies for Anatomical Dissection Regarded
as Control Subjects*
Nail Plasma RBC Urine Hair Kidney Spleen Liver Lung Bone
Subject Gender (ng/g) (ng/mL) (ng/mL) (ng/mL) (ng/g) (ng/g) (ng/g) (ng/g)
1 F 3.7 + 2.4 0.2 _+0.2 0.1 + 0.1 0.1+ 0.1 0.1+ 0.1
2 F 3.8 + 2.2 0.7 + 0.5 < 0.05 0.4 + 0.6 2.7 + 0.8
3 F 8.5 60.53 3.38 8.92 40.00 6.04
4 F 16.8 20.97 23.93 9.56 59.67 6.19
5 F 15.3 26.75 6.59 5.86 65.75 0.91
6 F 5.5 41.61 17.88 6.81 20.97 0.00
7 F 5.8 12.56 11.27 25.32 33.85 1.86
8 F 2.9 26.25 13.33 29.86 7.49 0.00
9 M 1.1 7.82 0.01 14.29 55.30 16.11
10 M 1.7 9.84 9.34 16.38 39.22 18.32
11 M 23.8 15.96 5.88 17.62 17.71 64.32
12 M 3.6 29.17 22.38 7.01 NDt 6.63
13 M 0.6 18.49 22.22 16.36 16.87 13.76
14 M 3.8 2.27 16.59 27.03 51.52 57.02
* Subjects 1 and 2: mean + SD, n = 3. Subjects 3-14: mean, n = 2.
~" Not determined.
Journal of Analytical Toxicology, Vol. 23, November/December 1999
with 0.5% HNO3and a decreasedsensitivitybecause of high salt of major matrix elements was very high, more than 98.5%.
contents. In GFAAS,HC104 may form a chloride of Ge that is The concentrations of Ge in hair and nail were high in all
easily volatilized in the furnace. The concentration of Fe in patients, but very low in controls, suggesting that these speci-
sample must be checked in MIP-MSwithout extractionbecausea mens were useful for biologicalmonitoring of Ge. Ge concentra-
molecular ion was observed at rn/z 72 with Fe solution at 10 tions in fingernailsand toenailswere ofthe same order in Patients
lJg/mLor higher.A molecular ion of ~Fe]~ havingthe same m/z A, B, and D. In the case of Patient C, Ge concentrations in finger-
72 was thought to have formed. Among biological specimens, nails were much lower than in hair, and Ge in toenails and in
attention must be paid to iron-rich samplessuch as erythrocytes, blood plasmawere very low, in the rangeof the controls. The dif-
liver, and spleen. ferencesin Ge concentrations in these specimensmay reflectthe
The sensitivity of GFAASwas lower than that of MIP-MS. remaining amounts or chemical speciesof Ge in the body. Hair
However, it was adequate for determining Ge concentrations in analyses gave useful information about Ge intake in the past.
specimens from patients in order to check the Ge accumulation Patient C stopped ingesting Ge compound about six months
in the body.This technique is necessary for the extraction pro- before sample collection.Assumingthat hair growth rate was 1.1
cess; however, it is valuable for practical use because GFAAS cm/month, Ge was continually excreted in hair after Ge intake
equipment is widelyavailable in many laboratories. was stopped. The excreted amount decreased gradually and
MIP-MSwas more sensitivethan GFAASand was comparableto reached undetectable levels during the three months prior to
ICP-MS (15). The extraction method was useful for determining sample collection.
Ge in specimenshavingvery low concentrationssuch as controls: Ge concentrations in specimens from controls were very
0.1--66 ng/g or ng/mL.This method would also be usefulfor iron- low, and their range was the same or a little lower compared
rich samples and for hard tissues such as bone and teeth, which with recent reports by ICP-MS (15). Schroeder and Balassa (1)
contain large amounts of Ca and P,becausethe removalefficiency reported Ge concentrations from 0.2 to 1.9 pg/mL in blood or
urine from normal persons. However,we found
one order of magnitude lower Ge amounts in
] lmklls blood and urine from three women who had no
history of using Ge tablets or Ge-enrichedfoods.
We also determined Ge concentrations in plasma
from 70 healthy adults working in the same fac-
tory to be 0.005 + 0.005 IJg/mL(17). The reasons
for the differences in Ge concentration are
It obscure, but we assumed that they would be
I related to the differencesin Ge content in foods
~g J and environments among countries or to the sen-
sitivity and specificityof analytical methods used.
70 71 72 73 74 75 78 77 70 71 72 73 74 75 76 7?
Therapeutic effects of organogermanium are
currently being examined. These analytical
Figure 3. Spectra of digested biological samples and standard solutions of Ge and other elements methods are expectedto be useful for determining
measured by MIP-MS in the rangefrom m/z 71 to 76. Identificaton: a, 10-ng/mL Ge standard solu- the accumulation state of Ge in the body in order
tion; b, erythrocytes (50 pL as original volume per milliliter measuring solution); c, 50 pL plasma; to prevent patientsfrom sufferingfrom side effects
d, 100 pL urine; e, 25 mg hair; f, 100 tJg/mL Fe; g, 100 pg/mL Ca; h, 100 pg/mL Mg; i, 100 pg/mL of Ge-containingdrugs.
Na; and j, 100 pg/mL K.
10 1. H.A. Schroeder and J.J. Balassa. Abnormal trace
metal in man: germanium. ]. Chron. Dis. 20:
.-. 8 211-224 (1967).
2. S.-T. Tao and P.M. Bolger. Hazard assessment of
germanium supplements. Reg. Toxicol. PharmacoL
4 3. O. Wada and M. Nagahashi. Germanium intoxica-
tion by foods for health. (In Japanese.) J. Jap. Med.
2 Assoc. 99:1929-1933 (1988).
4. G. Falkson and H. Falkson. Phase 2 trial of spiroger-
0 manium for treatment of advanced breast cancer.
1 2 3 4 5 6 7 8 9 Cancer Treat. Rep. 67:189-190 (1983).
Distance from scalp (cm) 5. F. Suzuki, R. Brutkiewicz, and R. Pollard. Ability of
sera from mice treated with Ge-132, an organic ger-
Figure4. Distribution of Ge in the hair sample from Patient C. The hair sample was cut into 1-cm manium compound, to inhibit experimental murine
pieces from the scalp, and Ge concentrations were determined. ascites tumors. Br. J. CancerS2:757-763 (1985).
6. T. Asaka, E. Nitta, T. Makifuchi, Y. Shibazaki,
Journalof Analytical Toxicology,Vol. 23, November/December1999
Y. Kitamura, H. Ohara, K. Matsushita, M. Takamori, Y. Takahashi, electrothermal atomization. Chem. Pharm. Bull. 28" 2687-2691
and A. Genda. Germanium intoxication with sensory ataxia. (1980).
J. NeuroL 5cis. 130:220-223 (1995). 13. C. Shleich and G. Henze. Trace analysis of germanium. Fresenius
7. A. Takeuchi, N. Yoshizawa, S. Oshima, T. Kubota, Y. Oshikawa, J. Anal Chem. 338:140-144 (1990).
Y. Akashi, T. Oda, H. Niwa, N. Imazeki, A. Seno, and Y. Fuse. 14. K. Jin, Y. Shibata, and M. Morita. Determination of germanium
Nephrotoxicity of germanium compounds: report of a case and species by hydride generation-inductively coupled argon mass spec-
review of the literature. Nephrom. 60:436-442 (1992). trometry. Anal. Chem. 63:986-989 (1991).
8. B. Hess, J. Raisin, Z. Zimmermann, F. Horber, S. Bajo, 15. J. Yoshinaga, M. Nakazawa, T. Suzuki, and M. Morita. Determination
A. Wyttenbach, and P. Jaeger. Tubulointerstitial nephropathy of trace elements in human liver and kidney by inductively coupled
persisting 20 months after discontinuation of chronic intake of plasma mass spectrometry. Anal 5ci. 5:355-358 (1989).
germanium lactate citrate. Am. J. Kidney Dis. 21 9548-552 (1993). 16. M. Chiba, A. Shinohara, Y. Inaba, S. Nakayama, H. Rinno, and
9. K. Asano, M. Yamano, K. Haruyama, E. Ilawa, K. Nakano, H. Koide. Two cases of patients taking germanium-containing food
M. Kurono, and O. Wada. Influence of propagermanium (SK-818) stuffs. (In Japanese.)Juntendo Med. J. 36:406-410 (1990).
on chemically induced renal lesions in rats. J. Toxicol. Sci. 19: 17. M. Chiba, A. Shinohara, K. Matsushita, H. Watanabe, and Y. Inaba.
131-143 (1994). Indices of lead-exposure in blood and urine of lead-exposedworkers
10. Y. Ishiwata, E. Suzuki, S. Yokochi, T. Otsuka, F. Tasawa, H. Usuda, and concentrations of major and trace elements and activities of
and T. Mitani. Studieson the antivial activity of propgermanium with SOD, GSH-Px and catalase in their blood. TohokuJ. Exp. Med. 178:
immunostimulating action. Arzneimittelforschung 44:357-361 49-62 (1996).
11. M. Anke and M. Glei. Handbook on Metals in Clinical and
Analytical Chemistry, H.G. Seiler,A. Sigel, and H. Sigel, Eds. Dekker,
New York, NY, 1993, pp 381-386.
12. Y. Mino, N. Ota, S. Sakao, and S. Shimomura. Determination of ger- Manuscript received September8, 1998;
manium in medicinal plants by atomic absorption spectrometry with revision received November 25, 1998.