Search

Background: Iron deficiency anemia (IDA) and thalassemia trait (TT) are the most common causes of microcytic hypochromic anemia (MHA) and are endemic in lower resource settings and rural areas with poor medical infrastructure. Accurate discrimination between IDA and TT is an essential issue for MHA patients. Although various discriminant formulas have been reported, distinguishing between IDA and TT is still a challenging problem due to the diversity of anemic populations., Methods: We retrospectively collected laboratory data from 798 MHA patients. High proportions of α-TT (43.33 %) and TT concomitant with IDA (TT&IDA) patients (14.04 %) were found among TT patients. Five machine learning (ML) approaches, including Liner SVC (L-SVC), support vector machine learning (SVM), Extreme gradient boosting (XGB), Logistic Regression (LR), and Random Forest (RF), were applied to develop a discriminant model. Performance was assessed and compared with six existing discriminant formulas., Results: The RF model was chosen as the discriminant algorithm, namely TT@MHA. TT@MHA was tested in an interlaboratory cohort with a sensitivity, specificity, accuracy, and AUC of 91.91 %, 91.00 %, 91.53 %, and 0.942, respectively. A webpage tool of TT@MHA (https://dxonline.deepwise.com/prediction/index.html?baseUrl=%2Fapi%2F&id=26408&topicName=undefined&from=share&platformType=wisdom) was developed to facilitate the healthcare providers in rural areas., Conclusion: The ML-based TT@MHA algorithm, with high sensitivity and specificity, could help discriminate TT patients from MHA patients, especially in populations with high proportions of α-TT patients and TT&IDA patients. Moreover, a user-friendly webpage tool for TT@MHA could facilitate healthcare providers in rural areas where advanced technologies are not accessible., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 Elsevier B.V. All rights reserved.)

Competing Interests: Declaration of competing interest The authors declare no conflict of interest.

During recent years, our understanding of the pathogenesis of inherited microcytic anemias has gained from the identification of several genes and proteins involved in systemic and cellular iron metabolism and heme syntheses. Numerous case reports illustrate that the implementation of these novel molecular discoveries in clinical practice has increased our understanding of the presentation, diagnosis, and management of these diseases. Integration of these insights into daily clinical practice will reduce delays in establishing a proper diagnosis, invasive and/or costly diagnostic tests, and unnecessary or even detrimental treatments. To assist the clinician, we developed evidence-based multidisciplinary guidelines on the management of rare microcytic anemias due to genetic disorders of iron metabolism and heme synthesis. These genetic disorders may present at all ages, and therefore these guidelines are relevant for pediatricians as well as clinicians who treat adults. This article summarizes these clinical practice guidelines and includes background on pathogenesis, conclusions, and recommendations and a diagnostic flowchart to facilitate using these guidelines in the clinical setting., (© 2014 by The American Society of Hematology.)

Iron deficiency is the most frequent cause of hypochromic microcytic anemia in children, but other causes, some of them requiring specific management, may be involved. Checking the iron-status is absolutely mandatory. When iron-status parameters are low, inadequate intake, malabsorption, blood loss, and abnormal iron utilization must be tested. In absence of iron deficiency, α- and β-globin and heme biosynthetic gene status must be checked. Assessing the iron stock level is difficult, because there is an overlap between the values observed in iron-replete and iron-deprived patients, so that at least 2 iron-status parameters must be below normal for diagnosing iron deficiency. Furthermore, inflammation may also mimic some characteristics of iron deficiency. Diagnosing iron deficiency leads to prescribing iron supplementation with follow-up at the end and 3 months after cessation of treatment. When iron stores are not replete at the end of treatment, compliance and dosage must be reevaluated and occult bleeding sought. The latter is also required when the iron store decreases 3 months after cessation of iron replacement., (Copyright © 2012. Published by Elsevier SAS.)

Many patients with breast cancer suffer from anaemia, as a consequence of the disease itself or its treatment. Anaemia has a negative impact on treatment outcome and overall survival, and affects the quality of life (QoL) of patients with cancer. Previously, cancer-related anaemia was treated with blood transfusion, but this is inconvenient, offers only temporary improvement in haemoglobin (Hb) level and is associated with several risks. Consequently, blood transfusion is usually reserved for patients with severe anaemia (Hb levels <8 g/dl). Recombinant human erythropoietin (epoetin) is an effective and convenient treatment for cancer-related anaemia without the risks associated with red blood cell transfusion. Epoetin therapy effectively increases Hb levels, thereby reducing the need for emergency blood transfusion and improving the QoL of patients with anaemia and breast cancer. Epoetin beta is also effective for the prevention of anaemia and reduction of transfusion requirements in patients with a high risk of developing anaemia during chemotherapy. With the increased use of dose-intensified chemotherapy in an attempt to improve response rates, administration of epoetin to prevent anaemia could potentially benefit many patients with breast cancer.

Current initiatives to reduce the high prevalence of nutritional iron deficiency have highlighted the need for reliable epidemiologic methods to assess iron status. The present report describes a method for estimating body iron based on the ratio of the serum transferrin receptor to serum ferritin. Analysis showed a single normal distribution of body iron stores in US men aged 20 to 65 years (mean +/- 1 SD, 9.82 +/- 2.82 mg/kg). A single normal distribution was also observed in pregnant Jamaican women (mean +/- 1 SD, 0.09 +/- 4.48 mg/kg). Distribution analysis in US women aged 20 to 45 years indicated 2 populations; 93% of women had body iron stores averaging 5.5 +/- 3.35 mg/kg (mean +/- 1 SD), whereas the remaining 7% of women had a mean tissue iron deficit of 3.87 +/- 3.23 mg/kg. Calculations of body iron in trials of iron supplementation in Jamaica and iron fortification in Vietnam demonstrated that the method can be used to calculate absorption of the added iron. Quantitative estimates of body iron greatly enhance the evaluation of iron status and the sensitivity of iron intervention trials in populations in which inflammation is uncommon or has been excluded by laboratory screening. The method is useful clinically for monitoring iron status in those who are highly susceptible to iron deficiency.

Despite the established utility of serum transferrin receptor (sTfR), serum ferritin, and the sTfR/log ferritin ratio (TfR-F Index) in the diagnosis of iron deficiency (ID) anemia, the numeric values of these parameters, which are indicative of subclinical ID, remain to be clearly defined. In this study, 65 apparently healthy nonanemic adults (22 men and 43 women) were treated with 3 months of oral iron supplementation to evaluate its effect on parameters reflecting iron status and to determine the prevalence of subclinical iron deficiency in apparently healthy adults. Significant supplementation-induced changes were observed in sTfR, ferritin, and TfR-F Index values in women, whereas in men, none of the studied parameters showed any significant change. Iron-deficient erythropoiesis (IDE) was not observed in men, but was found in 17 women (40%). Although individuals with a compromised iron status may be represented in substantial numbers in conventional reference populations, they can be readily identified using sTfR, ferritin, and TfR-F Index determinations., (Copyright 1998 by The American Society of Hematology.)

Anemia diagnosed early in pregnancy is associated with increased risks of low birth weight and preterm delivery. In several studies, the association between anemia and outcomes reversed direction during the third trimester; maternal anemia was no longer a risk factor for poor pregnancy outcomes. Camden study data were used to examine the probable cause of this observation. Maternal iron-deficiency anemia, diagnosed at entry to prenatal care, was associated with low dietary energy and iron, inadequate gestational gain, and twofold or greater increases in the risks of preterm delivery and low birth weight. During the third trimester, these associations (except with inadequate gestational gain) were no longer present. This reversal of risk status may be attributable to the poor predictive value of anemia and iron deficiency tests during the third trimester. However, the relationship between poor diet (with inadequate iron intake) and increased likelihood of preterm delivery persisted during the third trimester.

Human recombinant erythropoietin (rHu-Epo) is now extensively used in chronic renal failures; this treatment, resulting in a correction of the severe anemias seen in hemodialysed patients, may in turn lead to a resistance to rHu-Epo therapy by reason of the shortage of erythropoiesis factors, such as iron, vitamin B12 and folates. The utility of the red cell indices (MCV, MCH, RDW) for detection of early iron, folate and B12 deficiencies was studied in eighteen hemodialysed patients with end-stage renal failure treated with rHu-Epo; Microcytosis (MCV < 80 fl) was found ineffective in detecting iron deficiencies as well as macrocytosis (MCV > 100 fl) in folate and B12 deficiencies, partly due to the high incidence of associated iron and folate deficiencies. Lowered MCH (< 27 pg) was not more efficient than microcytosis in detecting early iron deficiencies. Increased RDW was the most sensitive feature for folate, iron and B12 deficiencies with respective sensitivities of 62.5%, 72% and 75%. The global specificity for detecting all deficiencies was 74%. However, high RDW values were not indicative of any type of deficiency; it may thus be concluded that RDW is a non expensive, non invasive and sensitive test, which allows a selection of hemodialysed patients treated with rHu-Epo for a complete investigation program, in order to detect early iron, B12 and folate deficiencies.

Recent studies have shown that the serum transferrin receptor is a sensitive, quantitative measure of tissue iron deficiency. This study was undertaken to determine the serum transferrin receptor's ability to distinguish iron-deficiency anemia from the anemia of chronic inflammation and to identify iron deficiency in patients with liver disease. The mean transferrin receptor level in 17 normal controls was 5.36 +/- 0.82 mg/L compared with 13.91 +/- 4.63 mg/L in 17 patients with iron-deficiency anemia (p less than 0.001). The mean serum receptor level was normal in all 20 patients with acute infection, including five with acute hepatitis, and was also normal in 8 of 10 anemic patients with chronic liver disease. Receptor levels were in the normal range in all but 4 of 41 patients with anemia of chronic disease. We conclude that unlike serum ferritin levels, which are disproportionately elevated in relation to iron stores in patients with inflammation or liver disease, the serum transferrin receptor level is not affected by these disorders and is therefore a reliable laboratory index of iron deficiency anemia.

The majority of anemias in the United States are characterized by low mean corpuscular volume and thus are classified as microcytic. Iron deficiency, chronic disease and thalassemia traits are the three leading causes of microcytic anemia. The true cause of anemia must always be sought so that the prevalence estimates of iron deficiency are accurate and so that appropriate treatment can be initiated for the anemic individual. In both the clinical setting and in surveys, the most frequent differential diagnosis of microcytic anemia will involve distinguishing between iron deficiency and chronic disease. Erythrocyte sedimentation rate (ESR), zeta-sedimentation rate (ZSR), and C-reactive protein (CRP) are elevated in a variety of diseases. These indicators may help differentiate the anemia of chronic disease from iron deficiency, so that iron deficiency is not overestimated in hospitalized and aged populations. The red cell distribution width (RDW) appears to be elevated to a greater extent in iron deficiency than in chronic disease or thalassemia traits. RDW and CRP are two of several indicators of iron status in the third National Health and Examination Survey (NHANES III).

Intervention strategies to combat iron deficiency anemia in developing countries may hasten the development of iron overload in patients with an inherited defect in hemoglobin synthesis. This risk could be diminished if there was a rapid and simple method available for detecting iron overload in population screening programs. We have developed such a method, which is in effect a semiquantitative ferritin measurement based on a modification of a two-site enzyme-linked immunoassay. The assay requires only 2 drops of whole blood and a total incubation time of 90 min. The procedure, which can readily distinguish iron deficiency from even a modest increase in storage iron, has a potentially wide application in settings where a prompt assessment of iron status is required.

The goal of detecting iron deficiency in children is to identify those whose Hb concentration will rise in response to treatment with iron. In a controlled treatment trial conducted among Eskimo children, we examined the effectiveness of various measures of iron nutrition in predicting a response to iron therapy (greater than 1.0 g/dl rise in Hb). A response was seen in 43%, and an additional 26% had an intermediate response (0.5 to 1.0 g/dl rise). When individual Hb values were expressed as SD scores of the Hb distribution of a reference population, a marked skew toward low scores was seen before treatment. After treatment, the distribution became more Gaussian, indicating that iron deficiency had been the major cause of anemia. Serum ferritin, transferrin saturation, and free erythrocyte protoporphyrin levels moved toward normal with treatment, however, none of the tests used was very effective in distinguishing individuals who would have a response t Hb from those who would not (sensitivities: 63 to 42%, specificities: 45 to 61%). Laboratory measures of iron nutrition were far more helpful in depicting the iron status of the population than they were in distinguishing iron-responsive from nonresponsive individuals.

We studied 4039 children who were 6 months to 12 yr of age to characterize developmental variations of serum ferritin and erythrocyte protoporphyrin. Age-related descriptive statistics were derived. The -2 SD value for serum ferritin was found to increase progressively from 12 to 21 micrograms/1 with increasing age, while the +2 SD value for erythrocyte protoporphyrin was found to decrease progressively from 65 to 42 micrograms/dl whole blood with increasing age. While the mean serum ferritin value was found to continue to rise throughout the first 12 yr of life, erythrocyte protoporphyrin values were highest at 1 to 2 yr of age, then fell to essentially constant levels after 4 to 6 yr of age. These relationships, as well as the linear relationship of increasing hematocrit and serum ferritin with increasing age, suggest that the rise of hematocrit with age, as previously observed, is associated with improving storage and availability of iron for heme synthesis.

The use of hemoglobin as a predictor of response to iron therapy, for screening, and for prevalence estimates was studied. An Fe supplementation trial was performed in Quito, Ecuador, in which 412 pregnant women were randomly assigned to treatment and control groups. Women in the treatment group received 390 mg ferrous sulfate/d for 2 mo. The prevalence of Fe deficiency as defined by response to therapy was found to be 60.8%. Sensitivity and specificity were calculated at various cutoff points of hemoglobin. The estimates of sensitivity and specificity allow for the use of hemoglobin in screening for Fe deficiency anemia and in the estimation of the prevalence in populations with characteristics similar to those found in the sample of pregnant women in Quito. Hemoglobin was shown to be a good predictor of response to Fe treatment and a good estimate of prevalence of Fe deficiency when prevalence is high.

169 hospital inpatients were studied to assess the association of blue sclerae with iron-deficiency anaemia. Three observers independently graded the signs of blue sclerae and mucosal pallor as absent, equivocal, definite, or striking. Blue sclerae were seen more often in patients with iron-deficiency anaemia (40/46, 87%) than in those with other anaemias (2/28, 7%; p less than 0.001) or without anaemia (5/95, 5.3%; p less than 0.001). The specificity of blue sclerae in iron-deficiency anaemia was 0.94 with a sensitivity of 0.87. By comparison, mucosal pallor was noted in only 30% of patients with iron-deficiency anaemia, with a specificity of 0.96 and a sensitivity of only 0.20 (p less than 0.001). The presence of blue sclerae was unaffected by age, sex, or colour of iris. Blue sclerae appear to be a good indicator of iron deficiency and should become a regular part of clinical examination.

Hematologic indices and iron balance data were obtained on 22 normal male volunteers who were subjected to a mean +/- SD phlebotomy of 164 +/- 34 ml whole blood/mo while living in a controlled environment. Over an average stay of 5 mo, volunteers did not develop anemia, but did display a reduction in iron stores that was quantitated by measurement of serum ferritin and iron balance. The percent saturation of transferrin and the usual erythrocyte parameters did not reflect changes in iron status. Loss of iron, which was calculated from quantitative phlebotomy and iron balance data, showed that a decrease of 1 ng of serum ferritin represented a loss of 4.5 +/- 5.3 mg of iron in 10 men whose initial serum ferritins were greater than 25 ng/ml, and 25.3 +/- 58.8 mg of iron in 7 men whose initial serum ferritins were less than 25 ng/ml. The period required for 3 volunteers who consumed a self-selected mixed diet at home to replace their depleted iron stores to prephlebotomy levels was about 4.5 mo. The sensitivity of serum ferritin as an index of iron stores was affirmed. In addition it was found that normal men who were consuming a mixed diet containing about 15 mg of iron daily and losing blood at a rate of 164 +/- 34 ml/mo did not increase their iron absorption sufficiently to compensate for the iron loss.

Tests to evaluate body iron stores were compared in patients with iron deficiency and the anemia of chronic disease. The serum ferritin assay separated these disorders in 20 of 22 patients. One discrepancy was explained by the concomitant association of both disorders. From this study and review of literature a low serum ferritin level is a good indication for iron therapy. The serum ferritin assay is a clinically useful test in lieu of bone marrow estimation of body iron stores to detect patients with iron deficiency. Total iron binding capacity levels when high-normal or elevated are sometimes helpful as a screening test in separating iron deficiency from the anemia of chronic disorders. Free erythrocyte protoporphyrin values were elevated in both conditions but were higher in iron deficiency than in the anemia of chronic disorders with considerable overlap of values. Urinary iron excretion with deferoxamine was not helpful in separating these disorders but is a useful test to establish iron overload. An elevated serum ferritin level is usually found with disease of iron overload but serum iron levels, deferoxamine iron excretion tests, and liver biopsy for estimation of iron stores are still beneficial diagnostic aids.

The diagnostic value of ferritin, transferrin saturation, and red blood cell zinc protoporphyrin for detecting iron depletion and predicting third-trimester anemia was studied in 87 women attending a private obstetrics clinic. A decline in ferritin and transferrin saturation and an increase in red blood cell zinc protoporphyrin levels were observed in third-trimester measurements when compared with those of the first trimester. Third-trimester anemia (hemoglobin less than 11.5 gm/dl or 115 gm/L) was detected in 13 (15%) women. Red blood cell zinc protoporphyrin was the only test that consistently demonstrated significantly different mean values between anemic and normal subgroups. The diagnostic sensitivity and predictive value of red blood cell zinc protoporphyrin for evaluating iron depletion and risk of anemia in pregnancy compared favorably to those of ferritin and transferrin saturation measurements. The operational simplicity and low cost of red blood cell zinc protoporphyrin measurements are additional characteristics that favor this procedure for office testing.

Laboratory results from the Second National Health and Nutrition Examination Survey (NHANES II) were used to define age-related changes in laboratory values that are related to the diagnosis of anemia and iron deficiency. Analyses included Hb, hematocrit, red blood cell count, red cell indices, iron, total iron-binding capacity, transferrin saturation, and erythrocyte protoporphyrin. Computation of median values and 95% ranges by age and sex for each laboratory test were performed on 15,093 subjects between 1 and 74 yr of age who had complete laboratory data on venous blood, after excluding those subjects with an abnormality in one or more of three other laboratory tests. Age-related changes in laboratory measurements, such as those described herein, must be taken into consideration in order to optimize the identification of individuals with anemia and iron deficiency.

One hundred sixty-one liver specimens from subjects who died acute traumatic deaths were studied. Of these, 127 came from male subjects and 34 from female subjects. The liver storage iron values of subjects in the persent study dying acutely from traumas and the differences found are in agreement with those expected according to the iron physiological needs and the prevalence of anemia reported by this cause in the world. This substantiates that the study of nonheme iron contents in subjects who died acute traumatic deaths is a good method for assessing iron nutritional status in a population. It is interesting to note the significant increase of hepatic storage iron found in men 61 to 70 years old, although it would be important to increase the number of cases to obtain more representative data. Liver storage iron in the Cuban population is not essentially different from that reported in the United States, Sweden, and Australia and the deficiency percentage is one of the lowest in Latin America.

Three formulae have been advocated for interpreting the blood-count when differentiating between iron deficiency and heterozygous thalassaemia: the discriminant function. M.C.V.-R.B.C.-(5 X Hb)-k, the M.C.V/R.B.C. ratio, and the (M.C.V.)2 X M.C.H. product. In a comparison of their diagnostic accuracy in microcytic adults from several countries, the discriminant function was the most satisfactory. 417 of 455 uncomplicated cases (91.6%) of iron deficiency, heterozygous alpha1- and beta- thalassaemia were correctly identified. The function was not able to distinguish heterozygous alpha2-thalassaemia from iron deficiency. A flow chart illustrates how the discriminant function can be used in haematological practice.

Hemoglobin, hematocrit, serum iron and total iron-binding capacity were measured in 38 apparently healthy children living in Quito (2:800 meters in altitude). One child was considered anemic according to the WHO reference adjusted to altitude (hemoglobin less than 12.3 g/dl) and 10 had transferrin saturation of less than 15%. Thirty children received orally 2 mg/kg per day of elemental iron for 3 mth. A significant increase in hemoglobin level, serum iron-concentration, and transferrin saturation, and a significant decrease in total iron-binding capacity were observed. Nine children increased their hemoglobin concentration by greater than 1 g/100 ml and could be considered as true anemics. At the end of the supplementation, only one child had a transferrin saturation coefficient of less than 15%. WHO references for hemoglobin adjusted to altitude seem to underestimate the prevalence of anemia for populations living at high altitudes.

The measurement of erythrocyte delta-aminolevulinic acid dehydratase (ALA-D) activity has been found to be a useful guide to lead exposure for screening purposes. The pH of the test should be well-controlled for adequate clinical interpretation. False-normal tests are rare even with iron-deficiency anemia or macrocytosis. Recent transfusion and recent bleeding or hemolysis with reticulocytosis tend to invalidate the test. ALA-D activity above 600 International Units effectively excludes significant lead exposure and eliminates the need for further testing. ALA-D activity is maximally depressed at a red cell lead concentration of 200 mug per 100 ml. When activity is moderately depressed, interpretation is difficult and the urinary excretion of lead after the administration of a chelating agent is usually required to differentiate the degree and significance of lead exposure.

The diagnostic usefulness in iron deficiency anemia of serum ferritin, red cell protoporphyrin (Epp), mean corpuscular volume, mean corpuscular hemoglobin (MCH), and transferrin saturation measurements has been studied in a population of 294 children aged 1 to 6 yr. Of the children studied 19% had hemoglobin below 11 g/dl. Iron deficiency, diagnosed by at least two abnormal independent laboratory parameters, was the cause of anemia in all except two cases. The Pearson correlation coefficient for hemoglobin was highest with MCH, followed in decreasing order of magnitude by MCV, Epp, transferrin saturation, and finally by ferritin. Sensitivity and specificity were highest for MCH and lowest for ferritin. Of anemic, iron deficient individuals 97 to 100% could be identified by low MCH, 88 to 100% by transferrin saturation, 66 to 83% by ferritin, and 61 to 74% by Epp. In contrast, only 0 to 6% of normal, nonanemic individuals had low MCH, 0 to 4% had high Epp, but 21 to 39% had low transferrin saturation and 25 to 39% had low ferritin. Although reduced serum ferritin in anemic individuals is good evidence of iron deficiency, a significant proportion of anemic iron-deficient patients is missed by this procedure rendering it less useful than other, less expensive laboratory methods.

To determine whether Hb criteria should be adjusted for obesity, the ranges and medians of Hb concentrations for 354 obese and 1894 nonobese, white children, aged 6 to 14 yr from HANES I were compared. Although median Hb concentrations average 0.125 g/dl higher in obese individuals with or without exclusion of iron-deficient individuals this slight difference is not significant (t tests p less than 0.1). It is suggested that while age, sex, altitude, and potentially, race are valid considerations when screening for anemia, a further adjustment for obesity appears unwarranted and would complicate efforts in optimizing screening for anemia.

In order to assess better the normal hemoglobin (Hb) range among adult males at high altitude, a hematological survey was completed in La Paz, Bolivia (3700 m). Two statistical methodologies were utilized to identify the anemic and polycythemic contributions to the Hb distribution of normal healthy men (n = 526). The Hb of normal men has a Gaussian distribution with mean and SD of 18.8 +/- 1.4 g/dl. Cutoff levels of 15.8 and 22.0 g/dl Hb were selected to identify the anemic and polycythemic subpopulations. Iron status in this high altitude population is similar to that of low altitude populations, and iron deficiency is the cause of most anemia found. An exercise test on a subsample of 56 men was utilized to validate the anemia cutoff level. Anemics as identified by the 15.8 g/dl Hb level showed a significant reduction in aerobic capacity as measured by PWC150 and estimated VO2max.

The coefficient of variation (CV) of red cell size, as measured by electronic red cell sizing (erythrography), was less than 14.0% in 20 normal subjects. In 22 of 25 patients with beta-thalassemia minor and microcytosis (mean corpuscular volume [MCV] less than 70 fl), CV was less than 14.0%; in the other 3, CV was 14.0%--14.9%. In 53 patients with iron deficiency anemia and MCV less than 70 fl, CV always was greater than 14.0%. In 7 patients with alpha-thalassemia minor and MCV less than 70 fl, CV was less than 14.0% in all 7. Among patients with microcytosis, erythrography appears to be an excellent technique for rapidly distinguishing between iron deficiency and alpha or beta thalassemia minor.

Serum carcinoembryonic antigen (C.E.A.) levels were measured in 381 undiagnosed patients who presented with clinical problems commonly associated with gastrointestinal malignancy. The results were compared with the final diagnosis after follow-up for up to 5 years to see whether C.E.A.-testing added any useful information. Of 307 patients presenting with upper gastrointestinal symptoms, lower gastrointestinal symptoms, or irom deficiency anaemia, C.E.A. levels greater than 20 ng/ml indicated malignancy in 5 but in 3 of these malignancy was also diagnosed after routine investigation. Of 74 patients presenting with obstructive jaundice, hepatomegaly, or abnormal liver function, malignancy was diagnosed in 38. In 9 of these patients the diagnosis of malignancy could otherwise have been reached only by laparotomy. The serum-C.E.A. thus reached only by laparotomy. The serum-C.E.A. thus seems to be of value in the assessment of liver disease but not in patients with gastric or colonic symptoms or iron-deficiency anaemia.