The progression of HCG during early pregnancy can be described only with frequent sampling of biological specimens. Our study of 37 naturally conceived pregnancies ending in live birth constitutes a relatively large and complete set of data for this purpose. Our results are based on collection of urine samples, which is a practical necessity when measuring HCG daily in healthy women. Even though blood may generally be preferred over urine for endocrinological assays, urine assays have proved to perform well for HCG measurements. Urinary concentrations of intact HCG are similar to serum concentrations (Wehmann and Nisula, 1981; Norman et al., 1987). We observe that doubling times for urinary HCG in early pregnancy are similar to those determined for serum measurements (Pittaway et al., 1985). In some circumstances, urine measures of HCG may even be superior to serum measures (Olsen et al., 2001). First morning voids of urine provide an integrated measure over time, thus smoothing possible pulsatile secretion of HCG by the trophoblast (Diaz-Cueto et al., 1994). The fact that commercial pregnancy tests and clinical kits rely on measurements of urinary HCG is yet another reason to be interested in the patterns of HCG excretion in urine during early pregnancy.Our measurements of HCG are based on well-characterized recognition sites for capture and detection antisera, as well as appropriate purified standards in each analyte assay (O'Connor et al., 1988, 1994). Each analyte being captured probably comprises a variety of isoforms. One example is hyperglycosylated HCG (H-HCG). This family of HCG variants has been the subject of much recent study (Weinans et al., 2000; Kovalevskaya et al., 2002), although the variants are not yet fully characterized. For example, it is not yet known how many H-HCG isoforms may be present in early pregnancy.The only forms of H-HCG that can be studied at present are those identified by the B-152 antibody. These forms are hyperglycosylated in the C-terminal region of the β subunit. Our capture antibody (B-109) does not bind near the C-terminal region, and therefore should not be affected by the structure of the C-terminal peptide. Based on epitope maps of HCG, this H-HCG should be fully captured by our assays for intact HCG and HCGβ (Birken et al., 2003).Final proof of this awaits the development of a true standard for H-HCG in pregnancy. What is currently regarded as ‘standard’ H-HCG has been derived from choriocarcinoma, and has a different molecular weight from the form found in pregnancy (Kovalevskaya et al., 2002). Also, the H-HCG derived from cancer cells is more likely to be ‘nicked’ HCG (another isoform variant) than the HCG in early pregnancy (Kovalevskaya et al., 2002). Once valid standards have been developed for H-HCG and its variants, more detailed studies of their patterns in early pregnancy will be possible.Our data describe the daily progression of several forms of HCG during the first 5 weeks after conception. In the earliest weeks of pregnancy, the HCGβcf has the lowest concentration of any of the measured HCG analytes. When HCGβcf has been generally regarded as the predominant form of urinary HCG in pregnancy (Kato and Braunstein, 1988; de Medeiros et al., 1992), this conclusion is based on samples collected later in pregnancy. Our data clearly show that the HCGβcf emerges as the dominant form only during the fifth week after conception. Conversely, HCGβ was relatively common in the first 3 weeks after implantation, and less common thereafter.Given that the urine samples in our study were stored and handled uniformly, it is unlikely that the relative changes between analytes that we observe during the first 6 weeks of early pregnancy could be due to degradation, disassociation or other changes attributable to storage. In particular, the relative gain in HCGβ over HCGβcf during the first few weeks of pregnancy is not easily explained by changes due to storage.Another unexpected finding was the high day-to-day variability of intact HCG in urine, both across women and within individual women. This variability was not due to physiological differences in urine concentration, and was confirmed using duplicate urine samples that had not been thawed previously. These abrupt changes cannot be explained by our current understanding of metabolism and urinary clearance (Wehman and Nisula, 1981). It is theoretically possible that an interfering factor in urine could disturb the binding to the specific capture antibody. No exogenous chemicals had been added to the specimens as preservatives, so interfering substances would have to be endogenous or the result of contamination of the urine sample during voiding or storage. These explanations seem unlikely given the sporadic nature of the fluctuations, and the fact that these fluctuations tended to occur on different days for different analytes for the same woman.Day-to-day changes in a woman's diet, personal habits or exposures might contribute to variability in urinary excretion of specific HCG analytes, although few examples are known. Levels of HCG (albeit later in pregnancy) have been reported to be decreased by cigarette smoking (Bernstein et al., 1989; Bremme et al., 1990). Only two of these 37 women reported smoking. We lack the detailed dietary or other daily information that would be needed to pursue such possibilities.Stability of these analytes over time must be considered in assessing our results, given that the urine specimens were 4–10 years old at the time of assay. The analytes we measured are among the most stable analytes of HCG. In a previous test of stability, we found that the concentration of intact HCG actually increased slightly over time, probably due to sublimation of water (Wilcox et al., 1985). Direct assessment of the stability of intact HCG, HCGβ and the HCGβcf showed no evidence of degradation under more stressful experimental conditions than experienced by our samples.If the unpredictable changes in urinary concentrations of intact HCG over successive days are a general occurrence in early pregnancy, there are clinical implications. Intact HCG has usually been regarded as the most bioactive of the several forms of HCG. More importantly, it provides the basis for many urinary assays for detection of pregnancy (Chard, 1992; Cole et al., 1993; Butler et al., 2001). To the degree that a pregnancy test relies solely on the detection of HCG, a single test could be falsely negative even a week or more after implantation (see ‘intact HCG’, Figure 2). This may be relevant to the performance of commercial urine-based pregnancy test kits (of which nearly 20 million are sold annually in the USA).Similarly, clinical measurements of HCG rate of rise based on urine assays might be distorted dramatically by the variability of intact HCG, especially if the rate of rise is determined on the basis of only two tests (as is often the case). These limitations of intact HCG could be compensated by expanding assays to detect additional forms of HCG. In our study, the combination assay appears to be extremely reliable and sensitive (Figure 1). [Our earlier Sepharose-IRMA was based on intact HCG and no doubt benefited from its cross-reactivity with HCGβ (Wilcox et al.., 1988).]There has been little discussion of sources of variation in urine HCG analyses (Lopata et al., 1982; Cole et al., 1993; Mishalani et al., 1994; Butler et al., 2001). It is possible that these fluctuations have biological significance, and that the physiological roles of intact HCG and its several forms may be more complex than previously suspected. For example, there could be paracrine or autocrine functions of HCG that evolve over the course of early pregnancy. If such mechanisms are operating, they are unlikely to be discovered without frequent sampling during the first weeks of pregnancy—ideally in serum as well as in urine, and using a battery of assays that describe multiple forms of HCG.In conclusion, our data show a gradually changing pattern of HCG forms excreted during early pregnancy, and reveal unexplained fluctuations in the day-to-day excretion of intact HCG (and to a lesser extent HCGβ). Variations over time among women and within women contribute to wide ranges of concentrations for intact HCG during the early weeks of pregnancy. The biological mechanisms of these variations and their functional significance are unknown at present. However, pregnancy tests that depend on the detection of urinary HCG may be more reliable and valid if they capture several HCG forms. This strategy is likely to provide a more sensitive assay, as well as one less subject to the apparent vagaries of concentration found with the intact HCG in early pregnancy. . Structure. Human chorionic gonadotropin is a glycoprotein composed of 237 amino acids with a molecular mass of 36.7 kDa, approximately 14.5 αhCG and 22.2kDa βhCG. It is heterodimeric, with an α (alpha) subunit identical to that of luteinizing hormone (LH), follicle-stimulating hormone (FSH), thyroid-stimulating hormone (TSH),. Introduction. HCG is the cornerstone of early pregnancy detection and monitoring. The most well characterized form of HCG is the intact heterodimer (intact HCG), composed of an α and β subunit..