Alpha-fetoprotein in plasma and serum of healthy adults: preanalytical, analytical and biological sources of variation and construction of age-dependent reference intervals

Alpha-fetoprotein (AFP) is a tumor marker for hepatomas and germ cell tumors, and the serum concentration has prognostic significance in other diseases. We examined the normal serum concentration of AFP in adults and sources of variation in the immunochemical variation of AFP. The serum concentration of the tumor marker alpha-fetoprotein (S-AFP) was log-normally distributed in 284 adult blood donors. S-AFP increased with age (p < 10(-7)), whereas no gender-related difference was found. Reference intervals (95-interpercentile) were constructed for persons < or =40 years (0.60-9.30 kIU/L) and >40 years (1.40 12.60 kIU/L). The concentration of AFP was significantly, albeit slightly, higher in serum than in plasma, whereas hemolysis, pretreatment with KCl and food intake did not influence S-AFP. S-AFP only changed 6% when measured twice 2 months apart (p=0.04). Three enzyme immunoassays, using three different anti-AFP monoclonal antibodies for detection, were compared and two assays gave S-AFP values significantly higher, 2.8% (p=0.03) and 19.0% (p<10(-4)), than the other assay. Thus, the choice of antibody may influence the result of immunochemical concentration determination. This can be explained by the existence of conformational variants of AFP with different antibody reactivities, and calls for careful standardization of monoclonal antibodies used in assays for AFP. With broad population reference ranges and slight intra-personal variation, the most effective reference range for S-AFP is previous values obtained in the same person.

Spontaneous regression of hepatocellular carcinoma

AIMS

To determine the clinical and histopathological features of a case of a spontaneous regression of hepatocellular carcinoma (HCC).

CASE DETAILS

HCC was found incidentally in a 73-year-old man during a laparotomy for evaluation of gastric retention. Despite no treatment being given, he improved gradually with no sign of tumour as evidenced by coeliac angiography 15 months later, as well as by explorative laparotomy after another 2 years. The patient died 15 years after the primary diagnosis of HCC, without known evidence of tumour recurrence. The patient's clinical records were reviewed, and paraffin-embedded liver tissue was re-evaluated. Both histological and immunohistochemical features were compatible with the diagnosis of a well differentiated HCC. Conceivable causes of the spontaneous regression of this and other reported cases are discussed, but the phenomenon remains enigmatic.

CONCLUSIONS

This case with a histologically proven HCC and a very long follow-up time confirms that spontaneous regression does occur. Since the case was found incidentally our report also implies that this may not be as rare as reported.

Circulating intercellular adhesion molecule 1 predicts non-specific elevation of alpha 1-fetoprotein

Molecules governing cellular interactions have been suggested to be involved in the spurious elevation of alpha 1-fetoprotein (AFP) in non-neoplastic liver disease. To explore this controversial issue, we measured AFP, circulating intercellular adhesion molecule 1 (cICAM-1), and common liver function tests in 111 patients (71 male, 40 female). Eighty-four patients had non-neoplastic chronic liver disease and 27 had hepatocellular carcinoma. The concentration of cICAM-1 was determined immunoenzymatically. In patients with non-neoplastic chronic liver disease, univariate analysis demonstrated a significant correlation between AFP and cholinesterase (R = -0.397, P < 0.001), aspartate aminotransferase (R = 0.421, P < 0.001), bilirubin (R = 0.231, P < 0.05) and cICAM-1 (R = 0.430, P < 0.001). Multivariate analysis among these variables and AFP indicated cICAM-1 to be the strongest independent predictor of AFP. We conclude that cICAM-1 compares favourably with liver function tests in predicting non-specific AFP variations in non-neoplastic chronic liver disease, suggesting a link between targeting of the inflammatory damage to the hepatocyte and development of neoplasia.

Serum alpha-fetoprotein and alcohol consumption

Fifty-nine persons, 23 chronic alcoholics and 36 normal healthy persons with a well described alcohol consumption, had the serum concentration of alpha-fetoprotein determined by a sensitive monoclonal immunofluorescent assay. A significant elevation in S-AFP was found in alcoholics, median 4.1 kIU/l as compared to 3.0 kIU/l in near-abstainers (< 12 g ethanol per day) (p < 0.02). This difference was not explained by differences in age. S-AFP correlated positively with age (p = 0.01). In non-alcoholics a borderline significant correlation with S-AFP was found with average daily alcohol consumption (self-reported) (p = 0.09) and a significant correlation with the serum concentration of carbohydrate-deficient transferrin (S-CDT) (p = 0.004). In 11 alcoholics 2 months of abstention from alcohol was accompanied by a median reduction of 21% in S-AFP (p < 10(-5)). In alcoholics, but not in social drinkers, S-AFP correlated with S-ASAT (p = 0.004). The increase of S-AFP with alcohol consumption may reflect reversible alcohol-induced liver affection.