Advertisement

Early detection of pregnancy after in vitro fertilization and embryo transfer with hyperglycosylated hCG versus Elecsys HCG+β assay: the hyperPOC study

  • Ernesto Bosch
    Correspondence
    Correspondence address. Ernesto Bosch, IVI-RMA Valencia, Plaza de la Policía Local, 3, PC, 46015 Valencia, Spain, Telephone: 0034 963060900
    Affiliations
    Human Reproduction Department, IVI-RMA, Plaza de la Policía Local, 3, PC, 46015 Valencia, Spain

    IVI Foundation – IIS La Fe. Avenida Fernando Abril Martorell, Torre 106 A, 7a planta, 46026, Valencia, Spain
    Search for articles by this author
  • Martin Hund
    Affiliations
    Roche Diagnostics Solutions, Roche Diagnostics International Ltd, Forrenstrasse 2, CH-6343, Rotkreuz, Switzerland
    Search for articles by this author
  • Reinhard van der Does
    Affiliations
    Medical Department, IST GmbH, Medardusring 120, 67112, Mannheim, Germany
    Search for articles by this author
  • Laura Caracena
    Affiliations
    Human Reproduction Department, IVI-RMA, Plaza de la Policía Local, 3, PC, 46015 Valencia, Spain

    IVI Foundation – IIS La Fe. Avenida Fernando Abril Martorell, Torre 106 A, 7a planta, 46026, Valencia, Spain
    Search for articles by this author
  • Author Footnotes
    † Affiliation at the time the study was conducted: Biostatistics, PROMETRIS GmbH, Mannheim, Germany
    Silke Ahlers
    Footnotes
    † Affiliation at the time the study was conducted: Biostatistics, PROMETRIS GmbH, Mannheim, Germany
    Affiliations
    Biostatistics Department, Excelya Germany GmbH, Soldnerstr 1, 68219 Mannheim, Germany
    Search for articles by this author
  • Elena Labarta
    Affiliations
    Human Reproduction Department, IVI-RMA, Plaza de la Policía Local, 3, PC, 46015 Valencia, Spain

    IVI Foundation – IIS La Fe. Avenida Fernando Abril Martorell, Torre 106 A, 7a planta, 46026, Valencia, Spain
    Search for articles by this author
  • Author Footnotes
    † Affiliation at the time the study was conducted: Biostatistics, PROMETRIS GmbH, Mannheim, Germany
Open AccessPublished:October 22, 2021DOI:https://doi.org/10.1016/j.rbmo.2021.10.007

      HIGHLIGHTS

      • hhCG and HCG+β measured at Day 4 after IVF-ET may aid early pregnancy detection.
      • hhCG and HCG+β had similar predictive performance when measured at Day 4, 7 and 11.
      • Performance of hhCG and HCG+β improved from Day 4 to 7 and Day 7 to 11 post-IVF-ET.
      • Superiority of hhCG over HCG+β was not shown.

      ABSTRACT

      Research Question

      Is measurement of hyperglycosylated human chorionic gonadotropic (hhCG) superior to beta-HCG (HCG+β) for early pregnancy detection after in vitro fertilization (IVF)-embryo transfer (ET)?

      Design

      Blood samples for the purpose of this study were collected on Day 4 (+1), 7 (+1) and 11 (+2) after ET from women aged 18–45 years undergoing first/second fresh/frozen IVF-ET cycles. Biochemical pregnancy was assessed on-site by hCG determination on Day 11; clinical pregnancy was assessed by ultrasound on Day 21 (+4/-3). Serum hhCG (immunochemiluminometric assay; Quest Diagnostics) and HCG+β (Elecsys® HCG+β assay, Roche Diagnostics) levels were measured. Performance of hhCG and HCG+β for predicting pregnancy was evaluated and cut-offs selected.

      Results

      In total, 155 women were enrolled and received IVF-ET. Area under the curve (AUC) (95% confidence interval) at Day 4 was not significantly different for hhCG (0.88; 0.83–0.94) and HCG+β (0.90; 0.84–0.95), as was predictive performance on Day 7 and 11, with higher AUC estimates versus Day 4. Applying cut-offs derived according to Youden's index at Day 4 (hhCG, 100 pg/ml; HCG+β, 1.30 mIU/ml), both biomarkers demonstrated high negative predictive values for ruling out pregnancy (hhCG, 83.8%; HCG+β, 82.8%) and high positive predictive values for ruling in pregnancy (hhCG, 89.0%; HCG+β, 84.9%) at Day 21. Diagnostic performance improved from Day 4 through 11.

      Conclusions

      Predictive performance for early pregnancy post-IVF-ET of Day 5 blastocysts was not significantly different for hhCG and HCG+β; hhCG superiority over HCG+β was not shown.

      KEYWORDS

      INTRODUCTION
      Infertility and its treatment with assisted reproductive technologies are psychologically and emotionally traumatic experiences for most couples (Malina and Pooley, 2017). One of the highest levels of treatment-associated stress occurs around the time of the human chorionic gonadotropin (hCG) pregnancy test (Yong et al., 2000), although stress and anxiety levels remain elevated throughout the cycle (Turner et al., 2013). For in vitro fertilization (IVF), the timeframe between embryo transfer (ET) and pregnancy testing currently ranges from 11 to 15 days, with this time interval thus representing a challenging period in patients’ lives. Interventions aimed at decreasing this waiting period are needed to reduce the associated stress of IVF. hCG is produced by syncytiotrophoblasts and is comprised of two highly glycosylated α and β protein subunits (Fournier, 2016; Guibourdenche et al., 2010; Kovalevskaya et al., 2007). Heterogeneity in either the protein or its carbohydrate content gives rise to different forms of hCG. Hyperglycosylated hCG (hhCG) is an isoform of hCG that contains more extensive and complex carbohydrate moieties. In contrast to hCG, hhCG is primarily synthesized by cytotrophoblasts (Guibourdenche et al., 2010; Kovalevskaya et al., 2007) and acts to facilitate blastocyst implantation into the uterine wall (Cole et al., 2006; Fournier, 2016; Guibourdenche et al., 2010; Handschuh et al., 2007). During the initial 3 weeks after embryo implantation, hhCG is the primary variant of hCG (Cole, 2007; Cole et al., 2003; Davies et al., 2003; Kovalevskaya et al., 2002; O'Connor et al., 1998; Sutton-Riley et al., 2006; Valmu et al., 2006).
      Testing for hCG and its free β-subunit (β-hCG) remain the gold standard for early pregnancy determinations, with testing around Day 13–15 after ET commonplace for the past two decades based on robust diagnostic performance (Chen et al., 1997; Porat et al., 2007). Testing for β-hCG on Day 11 after oocyte retrieval is also plausible, although the predictive value of a single β-hCG measurement for pregnancy outcome increases between Days 11 and 21 following oocyte retrieval (Ochsenkühn et al., 2009). Given its early synthesis in pregnancy, measurement of hhCG may permit earlier prediction of pregnancy during IVF than hCG/β-hCG, but there are important feasibility considerations. In particular, immunoassay antibodies vary in their relative specificities for hCG and hhCG; for example, the monoclonal antibody B152 has a specificity for hhCG of more than 99% relative to hCG (Cole, 2007; Sutton-Riley et al., 2006), whereas many other antibodies against hCG show significant cross-reactivity with hhCG (Cole et al., 2004). Furthermore, there is no established international standard reference material for hhCG and there is limited existing literature on its detection by different commercially available assays. The Elecsys® HCG+β assay (Roche Diagnostics International Ltd, Rotkreuz, Switzerland) was designed as an automated test and has been shown to detect hCG, hhCG, β-hCG and ‘nicked’ hCG that is missing the C-terminal peptide (Cole et al., 2011).
      Existing clinical data suggest the potential utility of assays that measure hhCG for earlier prediction of pregnancy than hCG-based tests. In a prospective blinded clinical trial enrolling 58 IVF-ET patients, a single serum or urine hhCG measurement identified pregnancies (both biochemical and clinical) on Day 6 post-ET with 100% sensitivity and specificity. The early detection of biochemical pregnancies was possible due to their association with lower hhCG values (Strom et al., 2012). Similarly, in a retrospective single-center evaluation of 112 women undergoing IVF-ET, a hhCG level >110 pg/ml on Day 9 after oocyte retrieval was 96% specific for ongoing pregnancy, yielding a positive predictive value (PPV) of 94% (Chuan et al., 2014). Moreover, Day 9 hhCG levels were more sensitive and had a larger area under the curve (AUC) than Day 9 hCG levels (0.87 versus 0.67, respectively). In contrast, diagnostic performance based on Day 16 measurements was similar for hhCG and hCG.
      We conducted a proof-of-concept study to determine if hhCG is superior to HCG+β measured using the Elecsys HCG+β assay for predicting clinical pregnancy when measured on Day 4 (primary objective), or Days 7 and 11 (secondary objectives) after transfer of a Day 5 blastocyst (equivalent to Days 9, 12 and 16 after oocyte retrieval, respectively).
      MATERIALS AND METHODS
      Study design and objectives
      This was a single-center (IVI-RMA, Valencia, Spain), prospective, proof-of-concept study to compare the predictive value of hhCG alone and Elecsys HCG+β assay (hereinafter called HCG+β) for diagnosing early pregnancy after IVF (clinical trial registration number NCT03184519 / 1701-VLC-011-EB [CIM RD002786]). Patients and investigators were blinded to hhCG assay results until the end of the study, and to HCG+β assay results on Days 4 and 7 after ET. HCG+β assay results were disclosed on Day 11 as part of routine clinical practice for pregnancy diagnosis. Enrollment occurred between June 2017 and June 2018.
      The primary objective was to determine whether serum hhCG measured on Day 4 after blastocyst ET was superior to HCG+β measured on Day 4 after ET in predicting clinical pregnancy on Day 21. Key secondary objectives were as follows: to determine whether hhCG measured on Days 7 and 11 was superior to HCG+β measured on the same days for predicting clinical pregnancy on Day 21; to identify cut-off values and compare the sensitivity and specificity, and the PPV and negative predictive value (NPV), of hhCG and HCG+β measured on Days 4, 7 and 11 for predicting clinical pregnancy on Day 21; to determine whether changes in hhCG on Day 7 versus Day 4 and Day 11 versus Day 7 were superior to corresponding changes in HCG+β in predicting clinical pregnancy on Day 21; to compare the sensitivity, specificity, PPV and NPV of changes in hhCG and HCG+β on Day 7 versus Day 4 and Day 11 versus Day 7 for predicting clinical pregnancy on Day 21. PPV and NPV were calculated based on pregnancies observed in the current study population rather than prevalence in the general population. Possible variance for days of measurement were +1 day for Days 4 and 7, +2 days for Day 11, +4/-3 days for Day 21.
      Study population
      Consecutive patients presenting for ET were screened for enrollment in the study. Eligible women were aged 18–45 years and were undergoing a first or second IVF-ET cycle. Cycles utilized either fresh or frozen, autologous, or donated oocytes. In all cases, one or two Day 5 blastocysts were transferred. Fresh ET in autologous cycles were performed after triggering with a single dose of
      250 µg recombinant hCG (OVITRELLE [choriogonadotropin alfa]; Merck KGaA, Darmstadt, Germany). Oocyte donation cycles were conducted after artificial endometrial preparation as described elsewhere (Labarta et al., 2017). Frozen ET were performed either with artificial endometrial preparation or within the natural cycle.
      Study procedures
      Demographic information, previous medical history and information relating to number of embryos transferred, fresh/frozen transfer, autologous or oocyte donor, and concomitant diseases/medications were collected at screening. hhCG and HCG+β were assessed on Days 4, 7 and 11 from serum samples (0.5 ml) collected from 6 ml of venous blood.
      Clinical pregnancy was confirmed on Day 21 after ET by transvaginal ultrasound only in women with a positive pregnancy test on Day 11. Women with a negative pregnancy test according to HCG+β finished the study on the day of the pregnancy test. Ultrasound for detection of a gestational sac, embryo, heartbeat, or ectopic pregnancy, was conducted at Day 21 and Day 28 (+4/-3). Pregnancy category (negative pregnancy, first trimester miscarriage, biochemical pregnancy, ongoing pregnancy) was documented at the final Day 65 (+7) visit or earlier in the case of premature termination of the study.
      Immunoassays
      Serum samples obtained for the assessment of hhCG concentrations were stored at -80°C before being shipped on dry ice to Quest Diagnostics (Quest Diagnostics, Nichols Institute, San Juan Capistrano, CA, USA) for analysis. hhCG was measured using an automated immunochemiluminometric assay that uses the hhCG-specific biotinylated antibody B152 as a coating antibody, and the hCG β-subunit-specific acridinium ester-conjugated antibody B207 as a labelled antibody (Pandian et al., 2003). The assay has a limit of detection of 0.1 µg/L and limit of quantification of 0.2 µg/L (coefficient of variation [CV] <20%), with an inter-assay CV of <7.5%. Serum samples obtained for the assessment of HCG+β concentrations were stored at -80°C before being shipped to Roche Diagnostics on dry ice for analysis using the Elecsys HCG+β assay (cobas e 411 analyzer), which measures hCG, hhCG, β-hCG and nicked hCG. The combination of biotinylated- and ruthenium-labelled monoclonal hCG-specific antibodies recognize the holo-hormone, ‘nicked’ forms of hCG, the β-core fragment and free β-subunit, with a functional assay sensitivity of <0.6 mIU/ml (i.e. the lowest analyte concentration that can be reproducibly measured with an intermediate precision CV of 20%). The Elecsys HCG+β assay has been standardized against World Health Organization guidelines (World Health Organization, 2007) and every reagent set has a barcoded label containing calibration information specific to that particular reagent lot (Roche Diagnostics GmbH, 2020).
      Statistical analysis
      The sample size was calculated for a comparison of the two tests, hhCG and HCG+β, by comparing the AUC of their receiver operating characteristic (ROC) curves, based on the methodology of Hanley and McNeil (1983) and a binormal model. Based on Chuan et al. (2014), AUCs of 0.79 for hhCG and 0.67 for HCG+β were assumed. The calculation further assumed that 45% of patients undergoing IVF present with clinical pregnancy at Day 21 and a correlation of 0.6 between the two tests, hhCG and HCG+β, for both the positive and negative groups. Allowing for 10% of patients being non-evaluable, a total of 152 patients were planned for enrollment to achieve 136 evaluable patients and 80% power for a one-sided z-test with a significance level α of 2.5%.
      Empirical AUCs for hhCG and HCG+β, and 95% confidence intervals (CIs) were estimated for combinations of day of pregnancy assessment and day of assessment of hhCG and HCG+β; differences between hhCG and HCG+β were compared using 95% CIs for the difference in AUCs and a pertinent chi-squared test. The observed presence or absence of pregnancy was used to determine test predictive performance in terms of sensitivity, specificity, PPV, and NPV after determination of cut-off thresholds for hhCG and HCG+β. Cut-offs for each test and endpoint were calculated using Youden's index (sensitivity + specificity -1), choosing the point for which the index value was maximal. The primary endpoint analysis used a formal comparison of both tests. All other tests (secondary endpoints) were purely exploratory and not for formal hypothesis testing.
      Patient baseline and demographic characteristics, and the course of biologic measurements over time were summarized with descriptive statistics. Continuous variables were summarized by mean, standard deviation, median, interquartile range (IQR, Q1–Q3), minimum, maximum, and number of available data. Categorical variables were summarized by absolute and relative frequencies and number of available data. Data were analyzed using SAS 9.4 software (Cary, NC, USA).
      RESULTS
      Study population
      A total of 155 patients were enrolled and received IVF-ET. Ninety patients discontinued the study early due to: negative pregnancy according to HCG+β test (n = 60), first trimester miscarriage (n = 14), lost to follow-up (n = 9), biochemical pregnancy (n = 5), or withdrawal of informed consent (n = 2) (Figure 1). Patient baseline characteristics and pregnancy history are summarized for all treated patients in Table 1. Briefly, patients had a median age of 39 years (minimum–maximum, 27–45) and a median body mass index of 22.3 kg/m2 (minimum–maximum, 18.0–39.0); 154/155 (99.4%) of patients were Caucasian/white. Eighty-six patients (55.5%) had a previous pregnancy with a median (minimum–maximum) of 1 (1–6) prior pregnancies and 1 (0–6) prior deliveries. Most patients had a single ET (n = 139; 89.7%). Frozen embryos were more commonly transferred than fresh embryos (n = 128; 82.6% versus n = 27; 17.4%). Over half of the procedures used autologous oocytes (n = 91; 58.7%) while the remainder used oocyte donation (n = 64; 41.3%).
      FIGURE 1
      FIGURE 1Patient disposition flowchart. *For four women, ongoing pregnancy was documented but the Day 65 visit not completed. HCG+β = human chorionic gonadotropin + β subunit.
      TABLE 1BASELINE CHARACTERISTICS AND PREGNANCY HISTORY
      CharacteristicN = 155
      Median age (IQR), years

      Min–max
      39 (35–41)

      27–45
      Categorized age, years, n (%)

      25–34

      35–40

      >40


      23 (14.8)

      84 (54.2)

      48 (31.0)
      Median weight (IQR), kg

      Min–max
      60 (55–68)

      45–110
      Median height (IQR), cm

      Min–max
      164 (160–168)

      149–183
      Median BMI (IQR), kg/m2

      Min–max
      22.3 (21–25)

      18.0–39.0
      History of any previous pregnancy, n (%)86 (55.5)
      Number of previous pregnancies, median (IQR)

      Min–max
      1 (1–2)

      1–6
      Number of previous deliveries, median (IQR)

      Min–max
      1 (0–1)

      0–6
      Complications during previous pregnancies, n (%)4 (4.7)
      BMI = body mass index; IQR = interquartile range; max = maximum; min = minimum.
      Predictive performance
      Using ROC analysis, AUC estimates for hhCG and HCG+β on Day 4 post-IVF-ET were 0.88 and 0.90, respectively (Figure 2). No significant difference was found between AUC values for these biomarkers on Day 4 (difference: -0.012 [95% CI -0.044–0.02], P = 0.46); superiority of hhCG over HCG+β could not be shown. Likewise, predictive performances were not significantly different for the two biomarkers based on Day 7 and Day 11 measurements, with higher AUC estimates than for Day 4, indicating improvement in predictive performance from Day 4 to Day 11 for both biomarkers (Figure 2). ROC curves for predicting clinical pregnancy at Day 21 based on change in hhCG or HCG+β from Day 4 to 7 and Day 7 to 11 post-IVF-ET (Figure 2) show that the predictive performance was not significantly different to the single measurement data, with hhCG and HCG+β showing an improved performance for the later versus the earlier time point.
      FIGURE 2
      FIGURE 2ROC curves for predicting clinical pregnancy at Day 21 based on hhCG and HCG+β measurements on (A) Day 4, (B) Day 7 and (C) Day 11, and change in measurements from (D) Day 4 to 7 and (E) Day 7 to 11 post IVF-ET. AUC = area under the curve; CI = confidence interval; ET = embryo transfer; hhCG = hyperglycosylated human chorionic gonadotropin; HCG+β = human chorionic gonadotropin + β subunit; IVF, in vitro fertilization; ROC = receiver operating characteristic.
      Diagnostic performance
      Cut-offs for hhCG and HCG+β for predicting clinical pregnancy were derived for each time point according to Youden's index and are presented in Table 2. For hhCG, Youden's index cut-offs were 100 pg/ml, 900 pg/ml and 2200 pg/ml at Days 4, 7 and 11, respectively. For HCG+β, Youden's index cut-offs at the corresponding time points were 1.30 mIU/ml, 23.40 mIU/ml and 54.10 mIU/ml, respectively. For both biomarkers, Youden's index cut-offs increased at each consecutive time point. Based on Day 4 post IVF-ET measurements, both hhCG and HCG+β demonstrated capability for diagnostic performance, with NPVs of 83.8% and 82.8% for ruling out pregnancy at Day 21 and PPVs of 89.0% and 84.9% for ruling in pregnancy at Day 21, respectively; corresponding sensitivity and specificity estimates were 86.9% and 86.4% for hhCG and 86.9% and 80.3% for HCG+β (Table 2). Diagnostic performance improved from Day 4 to Day 11; for example, for HCG+β, NPV increased from 82.8% to 96.7% and PPV from 84.9% to 92.0%; sensitivity also increased from 86.9% to 97.6% and specificity increased from 80.3% to 89.4% (Table 2). Good diagnostic performance was also observed based on changes in hhCG and HCG+β from Day 4 to 7 and Day 7 to 11 (Table 2).
      TABLE 2DIAGNOSTIC PERFORMANCE OF HHCG AND HCG+Β FOR PREDICTING CLINICAL PREGNANCY AT DAY 21 BASED ON MEASUREMENTS ON DAYS 4, 7 AND 11
      DayYouden's index cut-off
      Determined as the cut-off with the highest Youden's index.
      Pregnancy
      Number of events (presence of pregnancy on day 21) or non-events.
      Assay result
      Number of positive and negative assay results using the optimal cut-off.
      % (95% CI)AUC

      (95% CI)
      YesNo+-SensitivitySpecificityPPVNPV
      hhCG, pg/ml
      41008466826886.9

      (77.8–93.3)
      86.4

      (75.7–93.6)
      89.0

      (80.2–94.9)
      83.8

      (72.9–91.6)
      0.88

      (0.83–0.94)
      79008466906094.0

      (86.7–98.0)
      83.3

      (72.1–91.4)
      87.8

      (79.2–93.7)
      91.7

      (81.6–97.2)
      0.94

      (0.90–0.98)
      1122008366896098.8

      (89.8–99.2)
      89.4

      (79.4–95.6)
      92.1

      (84.1–96.7)
      98.3

      (86.5–99.0)
      0.98

      (0.96–100)
      Change from day 4 to 78008466896194.0

      (85.1–97.3)
      84.8

      (73.9–92.5)
      88.8

      (80.1–94.4)
      91.8

      (80.1–96.4)
      0.94

      (0.90–0.98)
      Change from day 7 to 1120008366856496.4

      (88.1–98.7)
      92.4

      (83.2–97.5)
      94.1

      (86.7–98.0)
      95.3

      (85.0–98.3)
      0.98

      (0.97–100)
      HCG+β, mIU/ml
      41.308466866486.9

      (73.6–90.6)
      80.3

      (70.4–90.2)
      84.9

      (75.8–92.2)
      82.8

      (67.9–88.3)
      0.90

      (0.84–0.95)
      723.408466846690.5

      (82.1–95.8)
      87.9

      (77.5–94.6)
      90.5

      (82.1–95.8)
      87.9

      (77.5–94.6)
      0.94

      (0.90–0.98)
      1154.108366886197.6

      (91.6–99.7)
      89.4

      (79.4–95.6)
      92.0

      (84.3–96.7)
      96.7

      (88.7–99.6)
      0.98

      (0.96–100)
      Change from day 4 to 718.308466896194.0

      (86.7–98.0)
      84.8

      (73.9–92.5)
      88.8

      (80.3–94.5)
      91.8

      (81.9–97.3)
      0.94

      (0.90–0.98)
      Change from day 7 to 1157.308366856496.4

      (89.8–99.2)
      92.4

      (83.2–97.5)
      94.1

      (86.8–98.1)
      95.3

      (86.9–99.0)
      0.98

      (0.97–100)
      AUC = area under the curve; CI = confidence interval; hhCG = hyperglycosylated human chorionic gonadotropin = HCG+β, human chorionic gonadotropin + β subunit; NPV = negative predictive value; PPV = positive predictive value.
      a Determined as the cut-off with the highest Youden's index.
      b Number of events (presence of pregnancy on day 21) or non-events.
      c Number of positive and negative assay results using the optimal cut-off.
      Laboratory findings
      From a median (IQR) of 100 (50–300) pg/ml at Day 4, hhCG levels increased to approximately 19-fold at Day 7 and to 89-fold at Day 11 for all patients who received IVF-ET (Figure 3; Supplementary Table 1). Similarly, HCG+β levels increased from a median (IQR) of 1.9 (0.1–5.9) mIU/ml at Day 4 to approximately 21-fold at Day 7 and to 118-fold at Day 11. Scatterplots showing the correlations between hhCG and HCG+β measured on Days 4, 7 and 11 are shown in Figure 4. Exploratory tests found a strong correlation between hhCG and HCG+β tests on Day 4 (Pearson's r = 0.98, Spearman's rho = 0.92; exploratory P < 0.0001 for both), on Day 7 (Pearson's r = 0.99, Spearman's rho = 0.99; exploratory P < 0.0001 for both) and on Day 11 (Pearson's r = 0.99, Spearman's rho = 0.99; exploratory P < 0.0001 for both).
      FIGURE 3
      FIGURE 3Boxplots showing changes in (A) hhCG and (B) HCG+β measurements following IVF-ET. ET = embryo transfer; hhCG = hyperglycosylated human chorionic gonadotropin; HCG+β = human chorionic gonadotropin + β subunit; IVF, in vitro fertilization.
      FIGURE 4
      FIGURE 4Scatterplot showing correlation between hhCG and HCG+β measurements on (A) Day 4, (B) Day 7 and (C) Day 11 post IVF-ET. ET = embryo transfer; hhCG = hyperglycosylated human chorionic gonadotropin; HCG+β = human chorionic gonadotropin + β subunit; IVF, in vitro fertilization.
      DISCUSSION
      This proof-of-concept study suggests that testing of hhCG and HCG+β has the potential to predict clinical pregnancy at Day 21 based on measurements as early as Day 4 following IVF-ET of Day 5 blastocysts using the globally available Elecsys HCG+β assay. In particular, the findings show that predictive and diagnostic performances for hhCG and HCG+β were not significantly different when measured at Days 4, 7 and 11 post-IVF-ET, with the performance of both assays improving from Day 4 to 7 and Day 7 to 11.
      Chuan et al. (2014) previously showed that hhCG was superior to hCG in detecting early pregnancy based on measurements at 9 days after oocyte retrieval; hhCG was also predictive of ongoing pregnancy confirmed by ultrasound between 8 and 9 weeks of gestation. hhCG has been used in some experimental conditions as a biomarker of clinical IVF pregnancies as early as 6 days after ET. For example, in a prospective blinded study of women undergoing IVF-ET, a single serum or urine hhCG measurement at 6 days post-ET provided 100% sensitivity and specificity for identification of pregnancies (Strom et al., 2012). Furthermore, biochemical pregnancies could be detected based on lower serum or urinary hhCG values, although the small size of the study (56 evaluable patients) and the use of all fresh embryos meant that the data were not definitive, and potential variations resulting from the use of frozen embryos could not be excluded (Strom et al., 2012).
      Cole et al. (2011) conducted a prospective, multicenter study that aimed to evaluate commercially available automated hCG tests at the time of initiation. The Elecsys HCG+β assay was shown to detect hhCG, demonstrating good detection (≥75% to ≤125% of standard concentration) of hCG (109%), hhCG (78%), nicked hCG (100%) and β-hCG (102%). In the present study, the 80.3–89.4% specificity of Elecsys HCG+β for predicting clinical pregnancy at Day 21 based on Day 4, 7 and 11 measurements was comparable to previously published data for hhCG (Cole et al., 2004). This sensitivity for hhCG potentially explains why the HCG+β test in this study yielded results that were not significantly different to the Quest Diagnostics hhCG assay for early detection of clinical pregnancy.
      The Elecsys HCG+β cut-offs identified in this study were all acceptably higher than the functional sensitivity limit (<0.6 mIU/mL) of the assay as stated by the manufacturer. However, diagnostic performance on Day 4 was of limited clinical value. The proposed Youden's index cut-off for the primary endpoint of predicting clinical pregnancy at Day 4 was 1.30 mIU/mL, providing a sensitivity of 86.9%, specificity of 80.3%, PPV of 84.9% and NPV of 82.8%. In practical terms, the Elecsys HCG+β at this threshold failed to detect 11 of 84 pregnancies and recorded 13 false-positive pregnancies. For comparison, at a Day 4 cut-off of 100 pg/ml, the Quest Diagnostics hhCG assay provided a sensitivity of 86.9%, a specificity of 86.4%, a PPV of 89.0% and a NPV of 83.8% for prediction of clinical pregnancy. In practical terms, the hhCG assay failed to detect 11 of 84 pregnancies and resulted in 9 false-positive pregnancies. When translated into clinical practice, these findings suggest that clinical decisions based on Day 4 results should be made with caution. This is especially true for hormone replacement therapy (HRT) cycles, in which withdrawal of treatment could lead to a pregnancy loss in false negative cases. These figures tend to improve on Day 7, with an increased NPV of 91.7% for hhCG, thus diminishing the likelihood of incorrect clinical decision-making. Still, caution in HRT cycles is recommended. Conversely, comparable PPVs on Days 4, 7 and 11 for hhCG (89.0%, 87.8% and 92.1%, respectively) allow for a confident clinical value. Chuan et al. (2014) reported that Day 9 hhCG at a cut-off of 81 pg/ml was 95% specific, yielding a PPV of 95%, with similar values for Day 9 hCG at a cut-off of 4 mIU/ml, but markedly lower sensitivity (24% versus 54%). The present cut-offs for hhCG and HCG+β, although associated with lower specificity, achieve a higher sensitivity.
      Turner et al. (2013) found that stress and anxiety levels remained elevated across the IVF-ET cycle in women presenting for both first and repeat IVF cycles. Therefore, earlier detection of pregnancy after ET (e.g. Day 4 compared with Day 11 or later) offers potential advantages for patients undergoing IVF procedures in terms of reducing the waiting period and associated stress and emotional impacts. Unlike hCG, hhCG facilitates trophoblast invasion in the first trimester of pregnancy, and thus targeting hhCG is potentially useful for predicting miscarriage and treating ectopic pregnancies or pregnancy complications, such as pre-eclampsia and gestational trophoblast diseases (Evans et al., 2015; Fournier, 2016; Sasaki et al., 2008). The Elecsys HCG+β assay has potential benefits compared with the hhCG assay, with the latter only available at Quest Diagnostics in California, USA. The Elecsys HCG+β assay, in contrast, is globally available and can be performed in-house with the potential to detect early pregnancy with high sensitivity. Moreover, with high diagnostic performance in the detection of hCG and its β-subunit, together with hhCG and its nicked variant, the Elecsys HCG+β assay moves the standard closer toward a ‘total hCG’ test.
      This study has a number of strengths and limitations. Patients and investigators were blinded to the results of assay determinations of pregnancy, which were only revealed after the end of the study. The patient population characteristics, including age of patients undergoing IVF-ET, was consistent with a real-world population, and the results are therefore generalizable to clinical practice. The study also employed a commercially available, automated system for measuring HCG+β, with the result that the findings can be readily repeated. All patients received a single dose of 250 µg recombinant hCG (OVITRELLE); although this increases the homogeneity of the study population for evaluation of the two biomarkers, it may have an impact on the reproducibility of our findings in different patient populations. The fact that the Elecsys HCG+β assay also detects hhCG may be considered a limitation. As a proof-of-concept study, the Elecsys HCG+β assay appeared to perform as well as the Quest Diagnostics hhCG assay but claims of equivalency are not possible since the study was designed to show superiority of hhCG; thus, it was not planned and powered to show equivalence between the two assays. Further investigation in an adequately powered cohort would be required to show equivalency between the two assays, and to assess the influence that factors such as age, BMI and transfer of fresh or frozen oocytes may have on diagnostic performance. Furthermore, this was a monocentric study; more robust results may be obtained from conducting analyses across multiple sites. Further evaluation is required to validate not only the test's performance against established criteria, but also to corroborate the findings in a larger sample size, across multiple sites, to analyze the accuracy of the determination on prediction of ongoing pregnancy.
      In conclusion, while this proof-of-concept study did not show that hhCG testing was superior to the Elecsys HCG+β assay, these data suggest that both hhCG and HCG+β testing have the potential to predict clinical pregnancy as early as 4 days after IVF-ET of Day 5 blastocysts. With earlier pregnancy detection, there is the opportunity to diminish the psychological and emotional impact of uncertainty around pregnancy outcomes.
      ACKNOWLEDGMENTS
      The authors thank Ana Belén Borgoñoz, Maria Albiach, Franciso Ruiz and Neus Amat for their assistance with patient enrollment.
      DISCLOSURES
      Conflict of interest
      Ernesto Bosch has received honoraria from Ferring B.V., Gedeon Richter, Merck and Roche; acted as a paid consultant for Ferring B.V., Merck, Gedeon Richter, Roche and Abbott; and reports research cooperation with Gedeon Richter. Martin Hund is an employee of Roche Diagnostics International Ltd and owns stocks/shares in F. Hoffmann-La Roche Ltd. Reinhard van der Does, Laura Caracena and Silke Ahlers declare no conflicts of interest. Elena Labarta serves on the speaker bureau and receives consulting fees for Ferring B.V., and Gedeon Richter; received grants from Ferring B.V.; and reports participation on an advisory board for Merck.
      Authors’ roles
      E.B. contributed to study design, patient selection, data interpretation, drafting of the publication and final approval of the version to be published. M.H. contributed to study design, authorship of the protocol, data interpretation, drafting of the publication, and final approval of the version to be published. R.v.d.D. contributed to study design, remote and on-site study monitoring, drafting of the publication, and final approval of the version to be published. L.C. contributed to patient selection and follow-up, drafting of the publication, and final approval of the version to be published. S.A. contributed to protocol authoring, data analysis, data interpretation, drafting of the publication, and final approval of the version to be published. E.L. contributed to patient selection and follow-up, data collection and interpretation, drafting of the publication, and final approval of the version to be published.
      Ethical approval
      The study protocol received ethics committee and institutional review board approval (CEI IVI Valencia; first approval allowing recruitment to commence was obtained on February 28, 2017). The study respected the ethical principles of the approving ethics committee and institutional review board, and the Declaration of Helsinki of 1964 and its later amendments. In addition, local laws and Good Clinical Practice guidelines were followed in its implementation. Study collaborators were adequately trained, and all patients provided their written informed consent.
      Data availability
      Qualified researchers may request access to individual patient level data through the clinical study data request platform (https://vivli.org/). Further details on Roche's criteria for eligible studies are available here: https://vivli.org/members/ourmembers/. For further details on Roche's Global Policy on the Sharing of Clinical Information and how to request access to related clinical study documents, see here: https://www.roche.com/research_and_development/who_we_are_how_we_work/clinical_trials/our_commitment_to_data_sharing.htm
      Funding
      This study was sponsored by Roche Diagnostics International Ltd (Rotkreuz, Switzerland). Third-party medical writing support for the development of this manuscript, under the direction of the authors, was provided by Chloe Fletcher, MSc, of Ashfield MedComms (Macclesfield, UK), an Ashfield Health Company, and Jesse Quigley Jones (on behalf of MIMS Hong Kong), and was funded by Roche Diagnostics International Ltd (Rotkreuz, Switzerland). COBAS, COBAS E and ELECSYS are trademarks of Roche. All other product names and trademarks are the property of their respective owners.
      REFERENCES
      Chen, C.D., Ho, H.N., Wu, M.Y., Chao, K.H., Chen, S.U., Yang, Y.S., 1997. Paired human chorionic gonadotrophin determinations for the prediction of pregnancy outcome in assisted reproduction. Hum. Reprod. 12, 2538–2541.
      Chuan, S., Homer, M., Pandian, R., Conway, D., Garzo, G., Yeo, L., Su, H.I., 2014. Hyperglycosylated human chorionic gonadotropin as an early predictor of pregnancy outcomes after in vitro fertilization. Fertil Steril. 101, 392–398.
      Cole, L.A., 2007. Hyperglycosylated hCG. Placenta. 28, 977–986.
      Cole, L.A., DuToit, S., Higgins, T.N., 2011. Total hCG tests. Clin. Chim. Acta. 412, 2216–2222.
      Cole, L.A., Khanlian, S.A., Riley, J.M., Butler S.A., 2006. Hyperglycosylated hCG in gestational implantation and in choriocarcinoma and testicular germ cell malignancy tumorigenesis. J. Reprod. Med. 51, 919–929.
      Cole, L.A., Khanlian, S.A., Sutton, J.M., Davies, S., Stephens, N.D., 2003. Hyperglycosylated hCG (invasive trophoblast antigen, ITA) a key antigen for early pregnancy detection. Clin. Biochem. 36, 647–655.
      Cole, L.A., Sutton, J.M., Higgins, T.N., Cembrowski, G.S., 2004. Between-method variation in human chorionic gonadotropin test results. Clin. Chem. 50, 874–882.
      Davies, S., Byrn, F., Cole, L.A., 2003. Human chorionic gonadotropin testing for early pregnancy viability and complications. Clin. Lab. Med. 23, 257–64, vii.
      Evans, J., Salamonsen, L.A., Menkhorst, E., Dimitriadis, E., 2015. Dynamic changes in hyperglycosylated human chorionic gonadotrophin throughout the first trimester of pregnancy and its role in early placentation. Hum. Reprod. 30, 1029–1038.
      Fournier, T., 2016. Human chorionic gonadotropin: different glycoforms and biological activity depending on its source of production. Ann. Endocrinol. (Paris). 77, 75–81.
      Guibourdenche, J., Handschuh, K., Tsatsaris, V., Gerbaud, P., Leguy, M.C., Muller, F., Evain-Brion, D., Fournier, T., 2010. Hyperglycosylated hCG is a marker of early human trophoblast invasion. J. Clin. Endocrinol. Metab. 95, E240–244.
      Handschuh, K., Guibourdenche, J., Tsatsaris, V., Guesnon, M., Laurendeau, I., Evain-Brion, D., Fournier, T., 2007. Human chorionic gonadotropin produced by the invasive trophoblast but not the villous trophoblast promotes cell invasion and is down-regulated by peroxisome proliferator-activated receptor-gamma. Endocrinology. 148, 5011–5019.
      Hanley, J.A., McNeil, B.J., 1983. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology. 148, 839–843.
      Kovalevskaya, G., Birken, S., Kakuma, T., Ozaki, N., Sauer, M., Lindheim, S., Cohen, M., Kelly, A., Schlatterer, J., O'Connor, J.F., 2002. Differential expression of human chorionic gonadotropin (hCG) glycosylation isoforms in failing and continuing pregnancies: preliminary characterization of the hyperglycosylated hCG epitope. J. Endocrinol. 172, 497–506.
      Kovalevskaya, G., Kakuma, T., Schlatterer, J., O'Connor, J.F., 2007. Hyperglycosylated HCG expression in pregnancy: cellular origin and clinical applications. Mol. Cell. Endocrinol. 260–262, 237–243.
      Labarta, E., Mariani, G., Holtmann, N., Celada, P., Remohí, J., Bosch, E., 2017. Low serum progesterone on the day of embryo transfer is associated with a diminished ongoing pregnancy rate in oocyte donation cycles after artificial endometrial preparation: a prospective study. Hum. Reprod. 32, 2437–2442.
      Malina, A., Pooley, J.A., 2017. Psychological consequences of IVF fertilization – review of research. Ann. Agric. Environ. Med. 24, 554–558.
      Ochsenkühn, R., Arzberger, A., von Schönfeldt, V., Engel, J., Thaler, C.J., Noss, U., 2009. Predictive value of early serum beta-hCG levels after single blastocyst transfer. Acta. Obstet. Gynecol. Scand. 88, 1382–1388.
      O'Connor, J.F., Ellish, N., Kakuma, T., Schlatterer, J., Kovalevskaya, G., 1998. Differential urinary gonadotrophin profiles in early pregnancy and early pregnancy loss. Prenat. Diagn. 18, 1232–1240.
      Pandian, R., Lu, J., Ossolinska-Plewnia, J., 2003. Fully automated chemiluminometric assay for hyperglycosylated human chorionic gonadotropin (invasive trophoblast antigen). Clin. Chem. 49, 808–810.
      Porat, S., Savchev, S., Bdolah, Y., Hurwitz, A., Haimov-Kochman, R., 2007. Early serum beta-human chorionic gonadotropin in pregnancies after in vitro fertilization: contribution of treatment variables and prediction of long-term pregnancy outcome. Fertil. Steril. 88, 82–89.
      Roche Diagnostics GmbH. Elecsys HCG+β method sheet v19. 2020.
      Sasaki, Y., Ladner, D.G., Cole, L.A., 2008. Hyperglycosylated human chorionic gonadotropin and the source of pregnancy failures. Fertil. Steril. 89, 1781–1786.
      Strom, C.M., Bonilla-Guererro, R., Zhang, K., Doody, K.J., Tourgeman, D., Alvero, R., Cedars, M.I., Crossley, B., Pandian, R., Sharma, R., Neidich, J., Salazar, D., 2012. The sensitivity and specificity of hyperglycosylated hCG (hhCG) levels to reliably diagnose clinical IVF pregnancies at 6 days following embryo transfer. J. Assist. Reprod. Genet. 29, 609–614.
      Sutton-Riley, J.M., Khanlian, S.A., Byrn, F.W., Cole, L.A., 2006. A single serum test for measuring early pregnancy outcome with high predictive value. Clin. Biochem. 39, 682–687.
      Turner, K., Reynolds-May, M.F., Zitek, E.M., Tisdale, R.L., Carlisle, A.B., Westphal, L.M., 2013. Stress and anxiety scores in first and repeat IVF cycles: a pilot study. PLOS One. 8, e63743.
      Valmu, L., Alfthan, H., Hotakainen, K., Birken, S., Stenman, U-H., 2006. Site-specific glycan analysis of human chorionic gonadotropin beta-subunit from malignancies and pregnancy by liquid chromatography—electrospray mass spectrometry. Glycobiology. 16, 1207–1218.
      World Health Organization. WHO international standard 4th IS chorionic gonadotropin, human. NIBSC code: 75/589. Instructions for use. 2007; 3: 1–3.
      Yong, P., Martin, C., Thong, J., 2000. A comparison of psychological functioning in women at different stages of in vitro fertilization treatment using the mean affect adjective check list. J. Assist. Reprod. Genet. 17, 553–556.

      Appendix. Supplementary materials

      Biography

      Ernesto Bosch was born in Philadelphia in 1968, graduated in 1992, and specialized in Obstetrics and Gynecology in 1997. He obtained his Ph.D. in 1999 and a Global Executive MBA in 2020. Currently, he is Clinic Director at Instituto Valenciano de Infertilidad in Valencia, where he has worked since 2000.
      KEY MESSAGE
      Predictive performance for early pregnancy after IVF-ET of Day 5 blastocysts was not significantly different for hhCG and HCG+β; hhCG superiority over HCG+β was not shown.