Recommendations on the measurement and use of the alcohol consumption biomarker CDT. A position paper from the IFCC Working Group on CDT standardisation.

BACKGROUND: Carbohydrate deficient transferrin (CDT) is a biomarker for excessive alcohol consumption utilized in clinical and forensic medicine and workplace testing. Previously, many different analytical methods for CDT were used and the measurand varied considerably, making direct comparison of test results difficult. To end this confusion, the IFCC established a working group on CDT standardisation (WG-CDT) which completed its tasks in 2017. METHODS: This IFCC position paper by the WG-CDT summarizes state of the art information about the measurand and the analytical methods and gives concise recommendations for its utilization. RESULTS: The results achieved by the CDT standardisation process led to accuracy improvements in national external quality assessment schemes over the years. A brief review of ROC based comparison studies with the traditional biomarkers (GGT, MCV, ALT and AST) discusses the bias resulting from inadequate study populations. In large groups of the general population the superior diagnostic performance of CDT is confirmed. CONCLUSION: The relationship between alcohol intake versus resulting CDT is discussed as well as the cutoff and measurement uncertainty. Concerning the application in practice, potential pitfalls are considered and recommendations handling both analytical and preanalytical caveats are given. Finally, some examples of serious misunderstandings in publications about CDT are addressed.

[Improvement of carbohydrate deficient transferrin measurement by capillary zone electrophoresis using immunosubtraction of immunoglobulins and transferrin]. / Intérêt de l'immunosoustraction des immunoglobulines et de la transferrine dans le dosage de la transferrine désialylée en électrophorèse capillaire.

CDT (Carbohydrate Deficient Transferrin) is considered as the most efficient biomarker of alcohol abuse available for routine use. Among the various methods developed for its measurement, capillary zone electrophoresis (CZE) on the multicapillary analyzer Capillarys2 provides high quality results at high throughput. However, the non CDT specific measurement of protein absorbance at 200 nm may bring abnormal profiles in samples from patients with high polyclonal immunoglobulin level or monoclonal component. We evaluated the automated immunosubtraction procedure developed by the manufacturer in 48 samples with abnormal electrophoretic profiles that potentially could interfere with CZE measurement of CDT. Elimination of the serum immunoglobulins raised the number of interpretable profiles from 19 (40%) to 37 (77%). The immunosubtraction procedure failed in samples with a monoclonal component present at a concentration > 60 g/L and in some samples harbouring a partially degraded C3 fraction. Six samples identified as genetic BC transferrin variants were also included in the study and submitted to an automated transferrin subtraction procedure to ascertain whether the additional peak were actually transferrin glycoforms. After treatment, two samples were classified as homozygote C for transferrin due to the persistence of one of the supposed transferrin peak. In conclusion, immunoglobulin and transferrin subtraction allow a better CDT measurement in most samples with interfering monoclonal components and avoid misclassification of suspected transferrin BC or CD variants.

[Nephelometry or turbidimetry for the determination of albumin, ApoA, CRP, haptoglobin, IgM and transthyretin: which choice?]. / Turbidimétrie ou néphélémétrie: quel choix pour les dosages de l'albumine, l'ApoA, la CRP, l'haptoglobine, l'IgM et la transthyrétine?

Nephelometry, which is considered as the reference method for serum proteins determination requires a specific equipment. The majority of protein determinations are therefore carried out on biochemistry automats using turbidimetry. The objective of a CNBH group (Collège national de biochimie des hôpitaux) was to compare nephelometry and turbidimetry for 7 automats: 2 nephelometers, the BN Prospec (Dade-Behring) and Immage (Beckman-Coulter) and 5 biochemistry systems using turbidimetry, the Integra and Modular (Roche Diagnostics), the LX20 (Beckman-Coulter), RXL (Dade-Behring) and AU (Olympus). The study was based on the determination of sera collections (albumin, ApoA, CRP, haptoglobin, IgM, transthyretin) of 140 samples each: 110 limpid samples and 30 samples called HLI (hemolytic, lipemic or icteric). Fifteen hospitals took part to this work. An ANOVA analysis on limpid samples and quality control sera concluded to an "automat" effect for the 6 tested proteins but did not show a "method" effect, (i.e. nephelometry versus turbidimetry). On the other hand, the transferability of the results was expected to be better and an effort on the choice of the antibodies and the standardization procedures should be made.

[Multisite validation of CDT measurement by the %CDT TIA and the Tina Quant %CDT kits]. / Validation multicentrique du dosage de la CDT par les méthodes %CDT TIA et Tina Quant %CDT.

The measurement of CDT (Carbohydrate Deficient Transferrin) is an essential biological tool in the diagnosis and follow-up of alcohol abuse. It is also employed as a marker of abstinence for the restitution of driving licences. However, the precision of measurement, and the between laboratory homogeneity of the results are still discussed. The ion exchange followed by immunodetermination of CDT is available in two products, the Tina Quant %CDT (Roche, Mannheim, Germany) and the %CDT TIA (Bio-Rad, Hercules, United States). This multicentre study was undertaken: 1) to evaluate the analytical characteristics of these kits and the homogeneity of the results from one laboratory to another, independently of the method used, 2) to validate the differences between the proposed normal values of both kits, 3) to study the possibility of using commercial control sera as external quality control. Four analytical systems were included in the study (Roche Modular/Hitachi 717, Beckman Coulter Immage and LX20, Dade Behring BNII). Determinations were carried out on pools of sera, commercial control sera, kit controls, and 30 serums of patients. These latter were also analyzed in capillary electrophoresis in order to establish correlations between the techniques. The calibrations were stable over one 2 weeks period. The repeatability of measurements spread out from 3,1% to 24,7%, for a mean value lower than 10%. The commercial control sera provided reliable results, with values adapted to a routine quality control use. The results of the Bio-Rad applications were lower by approximately 20% than those of the Roche application, which justifies the difference of the normal values (2,6% versus 3%), and an identical classification of the patients in at least 27 of the 30 samples. We conclude that the analytical quality of the compared techniques, even if it could be improved, is sufficient to guarantee a good reliability of the results. An external quality control could be proposed by using the control sera that we tested.

[An anicteric cholectasis and nephrotic syndrome: AL amyloïdosis in a 51 years-old man]. / Une cholestase anictérique associée à un syndrome néphrotique chez un homme de 51 ans: découverte d'une amylose AL.

Amyloidosis is a multiple-organ disease for which the diagnosis is often confusing and thereby delayed. Here, we present an archetypal case illustrating such difficulties. A 51 years-old man presented a mixed dyslipemia in November 2002, in June 2004 he has finally been diagnosed with a primary AL-amyloidosis. Within these two years, the arising of a non-icteric cholestasis and a nephrotic syndrome have triggered the search for a disease related to a multiple-organ protein deposition. Confirmation of the AL-amyloidosis was obtained through an histological examination, including direct immuno-fluorescence. Amyloidosis is a life threatening disease that need to be diagnosed at an early stage, in order to maximise the therapeutic expectations. The average survival after the diagnosis of AL-amyloidosis is 5% at 10 years. Often, treatments are initiated late in the course of the disease, at a time when organ lesion are constituted, severely affecting the prognosis.

Carbohydrate-deficient transferrin (CDT) determination by nephelometry using a commercial kit. Analytical and diagnostic aspects.

Carbohydrate-deficient transferrin (CDT) has been proposed as the most efficient marker of alcohol abuse. Absolute and relative concentrations of CDT were measured with a commercial assay (%CDTTIA from AXIS-Shield, Oslo, Norway) using rate nephelometry for transferrin determination. One hundred eighty-eight alcoholic patients (154 males, 34 females) and 132 control patients (113 males, 19 females) were included in the study. Within-run and day-to-day imprecision were 3.15% and 9.77%, respectively. The calibration curve was stable for more than 4 months with a shift below 5%. The commercial assay lacked sensitivity (Se = 0.48), but was highly specific (Sp = 0.98). Lowering the cut-off from 6% to 4.6% raised the sensitivity of the %CDTTIA test to 0.76 with a specificity of 0.90. We conclude that this adaptation to the Array Protein System (Beckman-Coulter) is suitable for routine use and offers precise results. It, however, requires an adaptation of the cut-off value for patients and of the target value for kit controls.

Increased plasma fibrinolysis and tissue-type plasminogen activator/tissue-type plasminogen activator inhibitor ratios after ethanol withdrawal in chronic alcoholics.

The effects of alcohol withdrawal on fibrinolysis were studied in 10 middle-aged male chronic alcoholics institutionalized for withdrawal therapy. All patients were sampled on admission [day 1 (D1)] and 21 days after alcohol withdrawal [day 22 (D22)]. The overall plasma fibrinolytic capacity (OFC) was assayed by measuring the ability of patient plasma to generate D-dimers from a standardized fibrin clot, and tissue-type plasminogen activator (t-PA) and t-PA inhibitor (PAI-1) levels were assayed together with serum cholesterol, triglyceride and cholesterol fractions. At D22, the OFC significantly increased in seven patients [D1 = 10 +/- 0.7 microg/h (mean +/- SD), D22 = 17 +/- 7.4 microg/h; P < 0.01], while t-PA and PAI-1 levels decreased in all patients but two (t-PA: D1 = 16.6 +/- 5 ng/ml, D22 = 10.2 +/- 3.8 ng/ml; P < 0.001; and PAI-1: D1 = 46 +/- 39 ng/ml, D22 = 21 +/- 28 ng/ml; P < 0.01). This study clearly demonstrates an increase in overall fibrinolytic activity after alcohol withdrawal, which is mainly due to a decrease in PAI-1 levels. These changes induced by alcohol abstinence might provide clear benefit by reducing the risk of thromboembolic events and particularly of stroke associated with elevated PAI-1 levels described in heavy drinkers.

[Peripheral markers, future perspectives]. / Marqueurs périphériques, perspectives d'avenir.

Four of the recently described peripheral markers of alcohol abuse have been reviewed. The acetaldehyde adducts allow to detect an alcohol abuse lasting for several weeks, even after a recent alcohol withdrawal. Inversely, 5-hydroxytryptophol (5-HTOL) reflects the alcohol consumption of the last 24 hours. Its detection is possible after the blood alcohol concentration has disappeared. Its measurement is run in urine samples, thus without invasive sampling. The hyaluronic acid and the activity of beta-hexosaminidase are markers of hepatobiliary alcohol induced disorders more than direct markers of alcohol intake. Acetaldehyde adducts could be used as markers of long term alcohol abuse, CDT as a marker of the recent alcohol abuse, and 5-HTOL the detection of alcohol abuse of the past day.

Objective diagnosis of alcohol abuse: compared values of carbohydrate-deficient transferrin (CDT), gamma-glutamyl transferase (GGT), and mean corpuscular volume (MCV).

BACKGROUND: Alcohol abuse is roughly twice as common as alcohol dependence. Subjects with alcohol problems are usually diagnosed only when medical complications are present. Therefore, both doctors and patients need a method for early diagnosis of alcohol abuse. METHODS: The mean corpuscular volume, gamma-glutamyl transferase, and carbohydrate-deficient transferrin in alcohol abusers, alcohol-dependent patients, and "nonalcohol hospital" individuals were compared. RESULTS: For objective diagnosis of alcohol abuse, we found a sensitivity of 24%, a specificity of 96%, and a global predictive value of 63% for mean corpuscular volume; a sensitivity of 42%, a specificity of 76%, and a global predictive value of 61% for gamma-glutamyl transferase; and a sensitivity of 67%, a specificity of 97%, and a global predictive value of 84% for carbohydrate-deficient transferrin. CONCLUSIONS: Carbohydrate-deficient transferrin proves to be the best marker of alcohol abuse. It allows objective detection so that therapeutic action can be started early, which is easier and more effective than in alcohol dependence.

[Carbohydrate deficient transferrin: what's new 20 years later?]. / La transferrine déficiente en hydrates de carbone: quoi de neuf 20 ans plus tard?

The clinical diagnosis of alcohol abuse has frequently to be supported by biological data. These data help the practitioner to establish the diagnosis and to monitor the alcohol withdrawal and possible relapses. They also can help to demonstrate to the patient the existence of a denied alcohol abuse. CDT determination has been proposed for more than twenty years, but commercial kits are available since only 1990. A consensus has to be established concerning a reference method and the definition of the transferrin isoforms that should be included in the CDT. CDT is nevertheless considered in many studies as the best marker of alcohol abuse now available. Its sensitivity is at least equal to that of gamma glutamyltransferase (GGT), and its specificity remains very high even in the presence of other pathologies connected or not to the alcohol abuse. CDT is, despite the lack of standardisation and the cost of its determination, a basic marker of alcohol abuse more particularly in the circumstances of lack of efficiency of the other markers.

Serum lipoprotein(a) [Lp(a)] in alcoholic men: effect of withdrawal.

This study examines the effect of alcohol withdrawal on lipoprotein(a) [Lp(a)] in 24 male, middle-aged chronic alcohol abusers admitted for withdrawal therapy. Serum concentration of Lp(a) was determined before and during the first 3 weeks of abstinence. The changes in three sialylated proteins [Lp(a), alpha 1-antitrypsin (alpha 1-AT), and haptoglobin (Hp)] and in desialylated transferrin (CDT) were also determined in 14 patients. After the 3 weeks of withdrawal therapy, the mean and median Lp(a) concentrations increased (p = 0.0001). The changes in Lp(a) levels were not related to the changes in dietary intake nor to the decrease in total HDL, HDL3, HDL2 cholesterol, Apo A-I, and Apo B. In the subgroup of 14 chronic alcohol abusers, Lp(a) levels increased parallel with Hp and alpha 1-AT, whereas CDT decreased. It is concluded that the impact of alcohol on sialylated proteins may be one of the mechanisms responsible for the increase in plasma Lp(a) after alcohol withdrawal.

Nephelometric determination of carbohydrate deficient transferrin.

We describe a technique for measuring carbohydrate-deficient transferrin (CDT) in serum. Serum transferrin fractions are separated by anion-exchange chromatography on microcolumns. Sialic acid-deficient transferrin fractions are collected in the eluate, and transferrin is then quantified by a rate-nephelometric technique. Imprecision (CV) was 4-5% within-run and 7-9% between runs (n = 15). Comparison with an isoelectric focusing-immunofixation method for transferrin index (x) yielded y = 761x + 7, Sy/x = 39 mg/L. Assay of sera from 90 abstainers or moderate consumers of alcohol showed that 81 (90%) had CDT concentrations between 30 and 70 mg/L. Among 74 alcoholics admitted to an alcohol treatment center, 54 (73%) had CDT > 70 mg/L, i.e., the diagnostic sensitivity was 73% at a specificity of 90% (area under receiver-operator characteristic curve = 0.891).

[Carbohydrate-deficient transferrin and liver diseases. Study of 94 patients]. / Transferrine déficiente en acide sialique et maladies du foie. Etude de 94 malades.

OBJECTIVES AND METHODS: Carbohydrate-deficient transferrin has been proposed as a marker of alcohol consumption. The aim of this study was to evaluate the accuracy of the carbohydrate-deficient transferrin serum level, measured by ion exchange chromatography followed by radioimmunoassay (Kit CDTect), for the diagnosis of excessive alcohol intake in patients with liver diseases. Ninety-four patients (68 men, 26 women, age 21-71 years), 42 with alcoholic liver diseases and 52 with non-alcoholic liver diseases, were studied. Twenty-six patients consumed > or = 40 g alcohol per day (mean alcohol intake: 84 +/- 52 g per day) and were considered to be excessive drinkers. RESULTS: The sensitivity of carbohydrate-deficient transferrin for the diagnosis of excessive alcohol intake was 35%, and the specificity was 91%. By pairing carbohydrate-deficient transferrin with other markers of alcohol consumption, the sensitivity of the association of carbohydrate-deficient transferrin and gammaglutamyl transpeptidase was 96%, and the specificity was 59%. CONCLUSION: In patients with liver diseases, carbohydrate-deficient transferrin is a specific marker of excessive alcohol intake but a lack of sensitivity may limit its use.

Plasma lipids and alcohol consumption in alcoholic men: effect of withdrawal.

This study examines the effect of alcohol withdrawal on plasma lipids and particularly on HDL-cholesterol subfractions, in 18 middle-aged, clinically healthy but chronically drinking men, institutionalized for withdrawal therapy. Plasma lipids, total HDL and HDL3-cholesterol, Apo A-I and Apo B were assayed before and after 30-86 days of abstinence. A 38% decrease in mean total HDL-cholesterol levels was observed after withdrawal therapy (P = 0.0002), and this was due mainly to a drop in HDL3-cholesterol concentrations (-43%, P = 0.0002). The decrease in HDL2-cholesterol concentrations was also significant (-21%, P < 0.005) but less marked. These results were not dependent on quantities of alcohol ingested before therapy, on duration of hospitalization and on changes in dietary fat intake or smoking habits. Apo A-I levels decreased (-39%, P = 0.0002) and the magnitude of the decrease after alcohol withdrawal was positively related to the duration of hospitalization. Apo B levels increased (+24%, P < 0.005). Among the anthropometric parameters, arm muscle area was significantly higher after alcohol withdrawal. The energy and macronutrient intakes did not significantly change during hospitalization. It is concluded that the modifications of HDL-cholesterol, HDL3-cholesterol, HDL2-cholesterol Apo A-I and Apo B values were induced by alcohol withdrawal in this population of chronic french alcoholics.

The mathematics of measurement: sensitivity, specificity and predictive value of available tests.

The correct clinical interpretation of any laboratory marker requires a good knowledge of its positive and negative predictive values, which are calculated from the sensitivity and the specificity of the test and from the prevalence of the disease. This means that we need a good assay technique in the case of biochemical tests, and a well defined reference sample (control subjects), with the ability to distinguish between a diseased and a healthy subject. In the field of alcohol abuse markers, each of these notions presents some degree of difficulty. There is no perfect solution, and a compromise has to be agreed upon.

Alcohol intoxication and sialic acid in erythrocyte membrane and in serum transferrin.

Microheterogeneity of serum transferrin as well as erythrocyte membrane sialic acid content were examined in alcoholic patients and healthy controls. Both the sialic acid content of erythrocyte membranes and of the circulating transferrin were significantly lower in alcoholic patients than in controls. A moderate daily ethanol intake (less than 80 g) allowed to observe a proportional relationship between alcohol intake and the carbohydrate deficient forms of transferrin, and also a correlation between alcohol intake and the membrane sialic acid content. This supports the hypothesis of ubiquitary alterations of glycosylations in connection to ethanol intoxication. Additional disturbances could explain the absence of correlations between membrane sialic acid, pattern of abnormal forms of serum transferrin, and alcohol intake in heavy alcoholic patients.

[Increase of serum gamma-glutamyl transferase activity in diabetic patients is not linked to diabetes itself]. / L'augmentation de l'activité sérique de la gamma-glutamyl transférase observée chez les patients diabétiques n'est pas liée au diabète lui-même.

In order to assess the relationship between increase of gamma-glutamyl transferase and diabetes, we measured serum gamma-glutamyl transferase in 48 male and 43 female insulin-dependent diabetics aged 20-60 years, 49 male and 41 female non insulin dependent diabetics aged 30-60 years, and in 114 male and 124 female controls (blood donors) aged 20-60 years. A1C haemoglobin was assayed in 124 diabetics. There is no increase of serum gamma-glutamyl transferase activity in insulin dependent diabetics of both sexes when compared with the controls. There is an increase of serum gamma-glutamyl transferase activity in non insulin dependent diabetics when compared either with controls (men p less than 0.02, women p less than 0.02) or with insulin dependent diabetics (men p less than 0.05, women p less than 0.02). There is no correlation between serum gamma-glutamyl transferase activity and fasting blood sugar (148 diabetics). In non insulin dependent diabetics of both sexes there is no correlation between A1C haemoglobin and serum gamma-glutamyl transferase activity. These results suggest that the increase observed in diabetics is not caused by diabetes itself, ie, by hyperglycemia, but by diabetes associated pathologies.

Fabrication of birefringent gratings using nonlinear polysilane thin films.

Birefringence is shown to be induced in polysilane thin films by the process of chain scission induced by nonlinear absorption at visible wavelengths. This writable birefringence is easy to control and can be used for the fabrication of birefringent gratings and other optical devices.