| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Endocrinology and Metabolism |
1 Department of Medicine and 2 General Clinical Research Center at Albert Einstein College of Medicine of Yeshiva University, Bronx, NY.
aAddress correspondence to this author at: Department of Medicine, Division of Endocrinology and Metabolism, Albert Einstein College of Medicine, 1300 Morris Park Ave., Room G47, Bronx, NY 10461. Fax 718-430-2446; e-mail dstein{at}aecom.yu.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: We evaluated a 2-dimensional liquid chromatography–mass spectrometry (2D LC/MS) IDA method. Sample preparation was by off-line solid-phase extraction, and C-peptide separation was performed on an Agilent 1100 2D LC system with a purification method based on high-pressure switching between 2 high-resolution reversed-phase columns. Because of the low fragmentation efficiency of C-peptide, multiple-reaction monitoring analysis was omitted and selective-ion monitoring mode was chosen for quantification. Native and isotope-labeled ([M+18] and [M+30]) C-peptides were monitored in the +3 state at m/z 1007.7, 1013.7, and 1017.7.
Results: The assay was linear (r2 = 0.9995), with a detection limit of 300 amole (1 pg) on column. Inter- and intraday CVs for C-peptide were 
2%. Comparison with an established polyclonal-based RIA showed high correlation (r = 0.964). Plasma concentrations of total C-peptide measured by RIA were consistently higher than by IDA LC/MS, consistent with the higher specificity of IDAs compared with immunoassays.
Conclusions: The 2D LC/MS IDA approach eliminates matrix effects, enhancing assay performance and reliability, and has a detection limit 100-fold lower than any previously reported LC/MS method. Isotope-labeled C-peptide(s) can be clearly differentiated from endogenous C-peptide by the difference in m/z ratio, so that both peptides can be quantified simultaneously. The method is highly precise, robust, and applicable to pharmacokinetic detection of plasma peptides.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Immunoassays for plasma and urinary C-peptide analysis are well established. Their specificity is questionable, however, because of the lack of congruence among RIAs and between RIA and ELISA (7)(8)(9)(10)(11), which may be attributable to differences in antibody specificities or matrix effects (8)(9)(10)(12). Because of their increased specificity, isotope-dilution assays (IDAs)1 are an alternative to immunoassays for measuring C-peptide. Peptide quantification by liquid chromatography–mass spectrometry (LC/MS) is based on detection of only those ions that represent the intact molecule (6)(7)(13)(14)(15)(16); it thus offers the possibility of absolute specificity. Previously published LC/MS methods for quantifying C-peptide in urine and plasma had low sensitivity and low sample throughput, and required complicated sample preparation and large sample volumes (6)(7). We describe a novel sample preparation technique and a 2-dimensional (2D) LC/MS IDA method developed to overcome these limitations.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
preparation of c-peptide tracer and calibrators
Human-sequence isotope-labeled C-peptides with relative molecular masses (Mr) of 3038 [M+18] and 3050 [M+30] were synthesized via incorporation of multiple fluorenylmethyloxycarbonyl-labeled amino acids containing either 13C or 15N labels on an ABI 433A Peptide Synthesizer by solid-phase chemistry. Peptides were purified by preparative reversed-phase (RP) HPLC. Synthetic C-peptides ([M+18] and [M+30]) for infusion or use as internal standard (IS) were quantified by nitrogen and amino acid analysis. The C-peptide IS ([M+30]; Mr 3050) was used to prepare stock solutions (9.8 µg/L) in 50 mmol/L phosphate buffer (pH 7) containing 2 g/L human serum albumin, 250 kilounits/L Trasylol, and 6 g/L m-cresol, and stored at –20 °C until thawed for use. Concentrations of stock solutions of Mr 3050 IS were verified against the WHO human C-peptide reference standard 84/510 (10) by both specific C-peptide RIA (Linco) and by LC/MS with good agreement (within 5%).
ms quantification of c-peptide
To generate a calibration curve, we dissolved C-peptides in dilution buffer [5 g/L human serum albumin, 5 g/L hydroxypropyl ß-cyclodextrin, 15 mmol/L sodium EDTA, 0.8 mol/L guanidine hydrochloride (GuHCl), 25 mmol/L potassium phosphate (pH 8.0), 25 mL/L acetonitrile, 1.5 mL/L trifluoroacetic acid (TFA)] and injected 50 µL directly on the column, representing 2–8000 pg of C-peptide. For the IDA calibration curve, we added 0–150 µL of 9.8 µg/L Mr 3050 IS to 300 µL of a quality-control human plasma stock (containing 500 kilounits/L Trasylol) and brought the sample volume up to 450 µL with dilution buffer. Samples underwent solid-phase extraction (SPE) as described below. We injected 50 µL of the resolubilized lyophilisate, representing 
60 µL of plasma and 400 pg of endogenous C-peptide on column. The C-peptide concentration in the quality-control stock was calculated as the mean of 20 consecutive measurements calibrated against our [M+30] IS, the concentration of which had been determined by nitrogen analysis and RIA. This C-peptide stock became our standard against which measurement errors were calculated.
isotope-dilution assay
IDAs are based on quantitative detection of an analyte signal by MS relative to a heavy-isotope–labeled IS that is added at a known concentration (6)(7)(17). The analyte and IS are of identical amino acid composition and exhibit identical physiochemical properties except that they differ in their m/z ratio (e.g., 1007.7 vs 1017.7 for C-peptide and the [M+30] IS). Because the signal intensity is linear for both analyte and standard, the plasma concentration can be calculated directly from the ratio of the signal responses.
human plasma samples for c-peptide
All human studies were approved by the Albert Einstein College of Medicine Committee on Clinical Investigations (Institutional Review Board). All participants provided written informed consent before participation. We obtained samples from healthy persons participating in studies monitoring the effect of nutrient stimulation on insulin secretion and C-peptide output, and as such, samples were obtained under basal and stimulated conditions. For analysis in independent protocols, some participants received an infusion of 15N-labeled [M+18] C-peptide during our study. Results of in vivo 15N C-peptide pharmacokinetics will be reported separately. Briefly, participants received an intravenous bolus containing 6.5 g/m2 glucose, followed by a variable infusion of 182 g/L glucose to clamp the plasma glucose concentration at 10 mmol/L (180 mg/dL). Additional blood samples were collected into EDTA-containing tubes supplemented with Trasylol (500 kilounits/L) and placed on ice; after centrifugation plasma was stored at –80 °C before LC/MS analysis.
immunoassays
Plasma C-peptide was quantified by an RIA for human C-peptide according to the manufacturer’s instructions (product no. HCP-20K; Linco). Both intra- and interassay CVs for this assay are reported to be <5%, and the standards are calibrated against the WHO C-peptide reference standard 85/510 (10).
sample preparation: spe procedure
Method.
C8 SPE cartridges (Sep-Pak Plus; Waters) were activated by washing with 4 mL of absolute methanol and then washed with 5 mL of 50 mL/L acetonitrile. After 0.3 mL of plasma was enriched with 0.1 mL of IS (9.8 µg/L [M+30] C-peptide) and 100 µL of 50 mL/L acetonitrile, it was vortex-mixed and loaded with a P1000 pipette on Sep-Pak Plus C8 cartridges. We performed the sample loading and the first wash step with a 1-mL adjustable- volume pipette (blue tip) through the outlet of the SPE cartridge (opposite direction to conventional sample loading). Cartridges were washed with 1 mL of 50 mL/L acetonitrile, followed by 3 mL of 200 mL/L acetonitrile containing 1 mL/L TFA. C-Peptide was then eluted with 3.6 mL of 350 mL/L acetonitrile in 25 mmol/L potassium phosphate (pH 8) into 4-mL NUNCTM CryoTube vials. The C-peptide–containing fraction was collected, freeze-dried, and stored at 4 °C until LC/MS analysis.
Before LC/MS analysis, we solubilized the freeze-dried extracts with 75 µL of 8 mol/L GuHCl for 20 min, then added 0.15 mL of 50 mL/L acetonitrile containing 3 mL/L TFA. Samples were filtered through a 0.45 µm Ultrafree-MC centrifugation filter (Millipore) and transferred to autosampler vials with 400-µL flat-bottomed inserts containing an additional 40 µL of 8 mol/L GuHCl. The autosampler injected 50 µL of sample into the first-dimension column. All samples were analyzed in duplicate. The total sample volume was sufficient for 5 injections.
Calculation of C-peptide recovery after SPE.
We added 100 µL of 9.8 µg/L IS ([M+30]) to 0.3 mL of human plasma. Samples (n = 5) were prepared according to the SPE procedure described above. To recovery control samples, we added 0.1 mL of 9.8 µg/L IS ([M+30]) directly into 3.5 mL of SPE elution buffer (acetonitrile–phosphate buffer), followed by drying. All samples were reconstituted and analyzed by LC/MS in duplicate. Recovery was calculated (as a percentage) as the ratio of the mean sample area to mean control area.
instruments
We used an 1100 series 2D LC system (Agilent), equipped with binary and capillary pumps and a column compartment with 2-position Rheodyne thermostabilized valves. LC/MS analysis was performed on SCIEX API 365 and API 4000 triple-quadrupole mass spectrometers (Applied Biosystems) equipped with a Turboionspray source. The LC experimental setup is shown in Fig. 1
.
|
lc/ms operating conditions
We performed first-dimension chromatography with a complex gradient elution on a Jupiter C5 column [bead size, 5 µm; pore size, 300 Å; 100 x 2 mm (i.d.)] from Phenomenex equipped with a 4 x 2 mm (i.d.) guard column (Table 1
). The first-dimension separation was by mobile phases composed of 1 mL/L TFA in water (mobile phase A) and acetonitrile (mobile phase B). The second-dimension separation involved a complex gradient on a Pursuit C18 column [bead size, 3 µm; pore size, 180 Å; 50 x 2 mm (i.d.)] from Varian. The second-dimension separation mobile phases consisted of 4 mL/L formic acid, 10 mL/L isopropyl alcohol, and 20 mL/L methanol in water (mobile phase A) and acetonitrile (mobile phase B). Both columns and a Rheodyne switching valve were operated at 40 °C.
|
We used the switching valve to transfer samples loaded on the first-dimension column and the peak-heart cut of the analyte to the second-dimension column and performed MS analysis of the analyte on the second-dimension column while the first column continued regeneration and equilibration (Table 1
). The total method time was 16 min. Additional details are available (18).
lc/ms detection and quantification
Because of the low fragmentation efficiency of the analyte, multiple-reaction monitoring was omitted, and we analyzed C-peptides by selective-ion monitoring electrospray MS in positive ionization mode. The molecular ion for C-peptide was monitored in the 3+ state at m/z 1007.7 for native C-peptide, 1013.7 for 15N-labeled C-peptide tracer ([M+18]), and at 1017.7 for the IS ([M+30]; Fig. 2
). Both instruments were controlled by AnalystTM software. Quantification was by area ratio of analyte vs IS. When 15N-labeled C-peptide (Mr 3038) was infused, total C-peptide was calculated based on the sum of the m/z 1007.7 and 1013.7 signals, corrected for a peptide-hydrate background of 6% of the 1007.7 signal subtracted from 1013.7 to account for water binding to endogenous C-peptide (19) (data not shown). No correction was necessary for water binding to the 15N-labeled C-peptide, as this would lead to a nominal Mr of 3056 (m/z 1019.7), outside of the mass window monitored. The signal-to-noise (S/N) ratio was calculated by the SD method as implemented in Analyst software.
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). The sample prepurified by C8 SPE was desalted and purified on the first-dimension column by TFA-containing mobile phases. The peak of interest was transferred by a Rheodyne column-switching valve into the second-dimension RP/MS column. This column was simultaneously preflushed, via a T-shaped mixing connector, with low-organic mobile phase to improve C-peptide focusing on the second column. The C-peptide peak refocused on the second-dimension column was eluted by an MS-compatible mobile phase and quantified by MS.
Samples injected on the first column contained 0.34 mol/L potassium phosphate. Basic phosphates slowly dissolve column silica, leading to column deterioration. To avoid this effect, we reconstituted samples in 50 mL/L acetonitrile containing 3 mL/L TFA, with a final pH of 3. This method approach demonstrated excellent ruggedness (see below).
choice of columns
The first dimension chosen was a Jupiter C5 column. We chose Phenomenex-brand LC columns on the basis of availability of short (length, 4-mm) guard cartridges, which had minimal effect on retention times when short analytical columns were used and could be replaced with minimal analyte retention time shift. Additional considerations were pore size; 300 Å silica has a smaller surface area and thus faster analyte elution (compared with matrices with smaller pore sizes), generates lower back pressure, and can be regenerated and reequilibrated more rapidly. The use of high-concentration GuHCl in the sample prevented on-column protein precipitation, which greatly increased column lifetime; we made more than 1000 injections of SPE plasma extracts before replacing this column.
For the second dimension, we chose a Pursuit C18 column, a high-resolution, fast, MS-compatible column from Varian. Because of its relatively large pore size, this column generates less backpressure than typical 100– 120 Å RP columns. It is important to keep overall backpressure of 2-dimensional methods as low as possible.
method validation
SPE C-peptide recovery.
The typical yield of C-peptide from plasma after SPE was 83.9% (range, 78%–89%; n = 5).
Linearity.
We tested initial linearity with a 12-point dilution curve. Absolute C-peptide signal area was linear over a 500-fold range (0.010 to 5 ng) on column (mean r2 = 0.9973; n = 3 experiments; Fig. 3A
). The assay was further calibrated over 6 points against the [M+30] IS by serial dilutions of the IS with human plasma. The linearity was even better when C-peptide was measured in the IDA vs the [M+30] IS, with an intercept not significantly different from zero (r2 = 0.9995; y = 0.99x + 0.002; Fig. 3B
).
|
Precision.
The precision data are summarized in Table 2
and in Table 1
of the Data Supplement that accompanies the online version of this article athttp://www.clinchem.org/content/vol52/issue5/ .
|
We added known amounts of [M+30] human C-peptide to human plasma to achieve final concentrations of 0.8 to 4.1 µg/L, consistent with the range of concentrations in fasting and stimulated plasma samples (see physiologic data below), and measured the concentrations by LC/MS. As shown in Table 1
of the online Data Supplement, recoveries ranged from 99% to 102% of expected. The relative and absolute errors demonstrated no concentration bias, with the former being <3%.
To evaluate precision and reproducibility, we prepared quality-control samples composed of human C-peptide in plasma and synthetic [M+30] isotope-labeled C-peptide (9.8 µg/L) at mean (SD) concentrations of 0.9 (0.01), 3.3 (0.1), and 14.9 (0.1) µg/L. We measured 5 sets of 6 replicates over 5–10 days on the API 4000, with 0.2–2.4 ng of endogenous C-peptide injected on the column. The assay was highly reproducible with both the inter- and intraday CVs <2% (Table 2
). Additional reproducibility data generated on the API365 mass spectrometer over a 3-month period with C-peptide and 15N-labeled C-peptide tracer quality-control standards dissolved in an albumin buffer gave interday CVs of 3.5% and 2.7%. These measurements extended from February to May, during which 9 °C seasonal temperature fluctuations occurred in the laboratory. We have previously observed that temperature fluctuations can be an important source of measurement instability if not properly controlled.
The limit of quantification (LOQ) of our LC/MS assay on the API4000 was 0.16 µg/L (54 pmol/L), equivalent to 10 pg (3 fmoles) on column, which we determined by use of the IUPAC definition of 10 times the SD of the background noise of a C-peptide–negative plasma matrix divided by the slope of the calibration graph (20). At this concentration, the imprecision was well within 10% of expected. The limit of detection and the true lower LOQ were likely even lower given the high S/N ratio (1 pg on column yielded a S/N ratio of 8), the possibility of injecting on column the equivalent of >60 µL of plasma, and the extremely low between-sample carryover (<0.1%). An example of a plasma sample with a C-peptide concentration near the LOQ is shown in Fig. 3C
. [M+30] C-peptide added to healthy human plasma at a nominal concentration of 0.17 µg/L, extracted, and measured by 2D LC/MS gave a S/N ratio of 50 with minimal background.
comparison of ria vs ida by ms
We used RIA and IDA to simultaneously analyze 190 plasma samples from 8 separate human physiology studies. Plasma C-peptide concentrations varied over a 20-fold range, from 0.4–1 µg/L basally to 8 µg/L after nutrient stimulation. The full data set is presented in Fig. 1A
of the online Data Supplement, with the regression line and the 95% prediction intervals. The results between methods were highly correlated (r = 0.964). The results obtained by immunoassay were 70% higher than those obtained by IDA. The regression equation (y = 1.66x – 0.41) had an intercept that was not significantly different from zero.
Although the absolute difference between IDA and RIA values was larger at higher C-peptide concentrations, the relative differences were independent of concentration (see Fig. 1B
of the online Data Supplement). In addition, we used the regression equation to calculate the theoretical C-peptide concentrations for RIA from the IDA data. A plot of the relative residuals between the actual and theoretical RIA values was approximately symmetric around zero at all concentrations, indicating that the difference was not concentration dependent (see Fig. 1C
of the online Data Supplement).
stimulation of endogenous insulin secretion
C-Peptide concentrations measured by 2D LC/MS IDA in the basal state as well as in response to hyperglycemic stimulation are shown in Fig. 4
. Basal concentrations were constant during the fasting state and increased in a physiologic fashion in response to intravenous glucose. Two peaks were observed, an initial sharp peak 2–5 min after initiation of hyperglycemia, and a broad second peak thereafter, representing first- and second-phase insulin secretion, respectively (2)(3)(4). The C-peptide pattern is consistent with normal C-peptide secretion rates (2)(3)(4).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
IDAs are used to measure metabolites (17), small drug molecules in biological fluids (23)(24), and large complex molecules such as peptide hormones (6)(7)(13)(14)(15)(16)(18). For reported C-peptide assays, the lowest amount measured was 30 fmoles on column, which was obtained via initial immunoprecipitation/SPE followed by off-line RP chromatography and direct MS detection (7); the detection limit was also 30 fmoles for ultrafiltration followed by RP LC/tandem MS (6). Our method, with a detection limit <300 amoles, has a limit of detection 100-fold lower than previous LC/MS assays and equivalent to immunoassays for C-peptide (limits of detection, 30–60 pmol/L for C-peptide RIA and ELISA; Linco Research; data on file), taking into account that maximum plasma volumes for such assays are fixed.
Although this sensitivity gain is partly attributable to advances associated with our API 4000, the most important variable is the decrease in background attributable to efficient 2-dimensional chromatography (Fig. 2
and Fig. 3C
) (18). Even small amounts of residual cations from plasma or TFA from the first-dimension chromatography can cause significant ion suppression (18). Tandem MS can differentiate C-peptide from contaminants (6), but its low fragmentation efficiency can lead to a substantial loss in signal (60-fold decrease) (18). The use of 2D LC/MS reduces matrix effects and increases overall assay sensitivity.
Initial prepurification may also lead to analyte loss, with recoveries of 60% for ultrafiltration (6) and 
80% for SPE [Ref. (7) and this report]. Our SPE procedure [modified from Ref. (7)] enabled simple and rapid sample preparation while preserving high recovery. Other advantages of our approach were as follows: (a) the use of only 0.3 mL of plasma avoided the necessity of diluting larger volumes of plasma; (b) use of 50 mL/L acetonitrile without TFA for the sample dilution, loading, and initial wash steps avoided problems caused by plasma treatment at acidic pH; (c) off-line purification by Sep-Pak Plus is more convenient than with the classic Sep-Pak because Sep-Pak Plus cartridges have smaller dead volume and are less expensive; (d) use of a 1-mL pipette for sample loading and the first wash step eliminated syringe use and reduced preparation costs and the probability of clogging the SPE inlet filter; (e) elution of bound peptide with 350 mL/L acetonitrile in 25 mmol/L potassium phosphate (pH 8) increased C-peptide purity in the eluted (collected) fraction by selectively decreasing peptide hydrophobicity; (f) this method requires <4 min per sample, provides sufficient material for quadruplicate injections, and generates cleaner samples, increasing column lifetime (and reducing cost per sample).
Strong adsorption of C-peptide (and insulin) to the surface of plastic and glass vials, which increases with increasing C-peptide purity, can decrease recovery (18). Thus, the higher sample purity obtained by elution with 350 mL/L instead of 500 mL/L acetonitrile led to a decrease in C-peptide recovery after SPE (data not shown). The addition of salts to the dried SPE extract and the high final concentrations of phosphate and GuHCl prevented C-peptide binding to the polypropylene filter housing and tubes; therefore, silanol-treated vials were unnecessary (data not shown), and the use of RP columns for both dimensions of LC eliminated problems from matrix effects at high salt concentrations (Fig. 2
and Fig. 3C
) (12)(18).
Although a triple-quadrupole mass spectrometer was available, we initially did not pursue sensitivity enhancement by tandem MS because we achieved highly efficient separation and eliminated matrix effects with the 2-dimensional method, with similar backgrounds for actual plasma samples and the solvent blank (Fig. 2
). We have since determined that no additional benefit would be derived from the use of tandem MS over single-quadrupole detection (unpublished data).
IDAs detect only the authentic intact analyte, and as such, represent the emerging "gold standard" for analyte detection and quantification. Numerous immunoassays for C-peptide have been reported that use single as well as sandwich antibody methods [reviewed in Refs. (8), (10)]. Although all commercially available assays have been standardized against WHO international reference preparation 84/510, there is a surprisingly large degree of inconsistency in results for different assays (7)(9), and we observed mean measured plasma concentrations that were 60% higher by RIA than by 2D LC/MS (see Fig. 1
in the online Data Supplement).
The strong correlation of our 2D LC/MS assay with RIA and the physiologic response of plasma C-peptide to stimulated insulin secretion demonstrate the suitability of our method, but MS and immunoassay methods differ, with values being higher for the latter in all cases. Our assay, which detects C-peptide at lower concentrations than an established immunoassay method, is consistent with the specificity of IDA for detecting only intact C-peptide. Although the RIA used had minimal sensitivity to intact proinsulin, other cross-reacting species might exist. A C-peptide with 2 basic amino acids at the COOH terminus may result from incomplete cleavage of proinsulin molecules (5). The C-peptide LysArg species has been found to make up 0%–10% of C-peptide concentrations in random samples (7). C-Peptide cleavage products are another likely source of C-peptide immunoreactivity (25)(26), and proinsulin cleavage products can account for up to 30% of total insulin in circulation (7)(27). It is currently unknown whether clinical immunoassays detect these species, although one commercial immunoassay clearly does (25).
C-Peptide catabolic products with an increased half-life relative to intact C-peptide could contribute to C-peptide immunoreactivity. Our results concerning C-peptide production in response to glucose stimulation are consistent with our assay monitoring only the processing of full proinsulin to C-peptide, but more data are required to elucidate C-peptide metabolism and the relationship between immunoreactive species and intact C-peptide detected by IDA.
In summary, we have developed a rugged, cost-effective 2D LC/MS method for measurement of C-peptide in human plasma that has a detection limit 100-fold lower than those of previously reported IDAs and is comparable in sensitivity to established immunoassays. Results correlated highly with an established immunoassay for C-peptide across the physiologic range. This method provides a new tool to study the production and clearance rates of endogenously produced peptides and hormones implicated in many disease states, including obesity and diabetes. With continued improvements in peptide labeling and MS instrumentation, IDA could provide new insights into the in vivo metabolism of secreted proteins.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  R. R. Little, C. L. Rohlfing, A. L. Tennill, R. W. Madsen, K. S. Polonsky, G. L. Myers, C. J. Greenbaum, J. P. Palmer, E. Rogatsky, and D. T. Stein Standardization of C-Peptide Measurements Clin. Chem., June 1, 2008; 54(6): 1023 - 1026. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.-M. Wiedmeyer, K. S. Polonsky, G. L. Myers, R. R. Little, C. J. Greenbaum, D. E. Goldstein, and J. P. Palmer International Comparison of C-Peptide Measurements Clin. Chem., April 1, 2007; 53(4): 784 - 787. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
