Pregnancy outcomes of patients with low serum β-hCG
A total of 312 patients had serum β-hCG levels <300 mIU/ml 14 days after blastocyst transfer, among which, 18.6% were live births, 47.4% were early miscarriages, 22.8% were biochemical pregnancies and 9.6% were ectopic pregnancies. Rates of biochemical pregnancy loss, ectopic pregnancy, early and late miscarriages, live birth were comparable between the <38 years group, and the > 38 years group (Table 1). Among the 241 clinical pregnancies, 225 (93.4%) were singletons and 16 (6.6%) were twins (i.e., 9 monozygotic twins and 7 dizygotic twins). The rate of live birth was 24.9% (56/225) in singletons and 12.5% (2/16) in twins. The lower limits of serum-borne β-hCG levels were 64.9 mIU/ml for singleton live births, 145.1 mIU/ml for twin live births, 15.3 mIU/ml for early miscarriage, and 5.3 mIU/ml for ectopic pregnancies.
In sum, 164 patients had another serum β-hCG test at 2-24 days (mean: 6.75 days) after the initial measurement, among which, 133 had increased β-hCG levels and 31 had decreased values. For patients with decreased β-hCG levels, 96.8% (30/31) were biochemical pregnancy loss. The only patient, although presenting with declined β-hCG levels from 133.4 to 64.5 mIU/ml, eventually developed into an ectopic pregnancy.
Table 1 Pregnancy outcomes of patients with serum β-hCG level < 300 mIU/ml 14 days after blastocyst transfer
Pregnancy outcomes
|
<38 years
|
≥38 years
|
Total
|
P
|
%(n)
|
%(n)
|
%(n)
|
Female age (years)
|
31(28,34)
|
39(38,41)
|
|
1.13E-46 a
|
AMH (ng/ml)
|
5.83(3.56,9.34)
|
3.64(1.84,6.00)
|
|
7.45E-5
|
BMI (kg/m2)
|
21.9±3.2
|
23.6±3.7
|
|
0.001
|
Days of embryos transfer %(n)
|
|
|
|
0.645
|
5
|
68.2(176/258)
|
55.6(30/54)
|
|
|
6
|
29.5 (76/258)
|
44.4(24/54)
|
|
|
both 5&6
|
2.3 ( 6/258)
|
0( 0/54)
|
|
|
No. of embryos transferred % (n)
|
|
|
|
0.954
|
1
|
43.0(111/258)
|
42.6(23/54)
|
|
|
2
|
57.0(147/258)
|
57.4(31/54)
|
|
|
Biochemical pregnancy loss
|
22.5(58)
|
24.1(13)
|
22.8(71)
|
0.800
|
Live birth % (n)
|
19.0(49)
|
16.6(9)
|
18.6(58)
|
0.690
|
Ectopic pregnancy % (n)
|
10.1(26)
|
7.4(4)
|
9.6(30)
|
0.545
|
Early Miscarriage % (n)
|
46.9(121)
|
50.0(27)
|
47.4(148)
|
0.678
|
Late miscarriage % (n)
|
1.5 (4)
|
1.9 (1)
|
1.6 (5)
|
1.000 b
|
a P values meant to compare differences between the groups of <38 years and ≥38 years .b Fisher exact test were used.
Pregnancy outcomes of patients with different β-hCG intervals
For patients with β-hCG levels of 5-50 mIU/ml, no live birth occurred. Most of them were biochemical pregnancies (77.8%) and the remainder were early miscarriages (13.9%) and ectopic pregnancies (8.3%). Among patients with β-hCG levels of 51-100 mIU/ml, it was found that 55.8% were early miscarriages, 25.0% were biochemical pregnancies, and only 7.7% were live births. For patients with β-hCG levels of 101-200 and 201-299 mIU/ml, the likelihood of live births was about 1/4 (i.e., 23.7% and 24.5%, respectively) and the probability of early miscarriage was about 1/2 (i.e., 50% and 51.9% respectively; Table 2).
Table 2 Pregnancy outcomes of patients with different β-hCG levels 14 days after blastocyst transfer
HCG level mIU/ml
|
5-50
|
51-100
|
101-200
|
201-299
|
P
|
Female age (years)
|
33.0±4.5
|
32.2±4.2
|
32.2±4.8
|
32.6±5.2
|
0.835
|
AMH (ng/ml)
|
6.18±4.37
|
6.79±4.30
|
6.80±4.78
|
6.60±4.89
|
0.922
|
BMI (kg/m2)
|
21.6±2.9
|
22.5±3.0
|
21.9±3.2
|
22.6±3.8
|
0.269
|
Days of embryos transfer %(n)
|
|
|
|
|
0.356
|
5
|
63.6(21/33)
|
75.9(41/54)
|
58.8(70/119)
|
69.8(74/106)
|
|
6
|
33.3(11/33)
|
24.1(13/54)
|
38.7(46/119)
|
28.3(30/106)
|
|
both 5&6
|
3.0(1/33)
|
0(0/54)
|
2.5(3/119)
|
1.9(2/106)
|
|
No. of embryos transferred % (n)
|
|
|
|
|
0.609
|
1
|
45.5(15/33)
|
44.4(24/54)
|
46.2(55/119)
|
37.7(40/106)
|
|
2
|
54.5(18/33)
|
55.6(30/54)
|
53.8(64/119)
|
62.3(66/106)
|
|
Serum β-hCG(mIU/ml)
|
34(29,41)
|
80(61,90)
|
145 (127,174)
|
244(218,244)
|
1.93E-149
|
Biochemical pregnancy loss %(n)
|
75.8(25/33)
|
29.6(16/54)
|
13.4(16/119)
|
13.2(14/106)
|
3.81E-14
|
Early miscarriage %(n)
|
15.2(5/33)
|
53.7(29/54)
|
49.6(59/119)
|
51.9(55/106)
|
0.001
|
Ectopic pregnancy %(n)
|
9.0(3/33)
|
7.4(4/54)
|
11.8(14/119)
|
8.5(9/106)
|
0.827a
|
Late miscarriage %(n)
|
0.0(0/33)
|
1.9(1/54)
|
1.7(2/119)
|
1.9(2/106)
|
1.000a
|
Live birth %(n)
|
0.0(0/33)
|
7.4(4/54)
|
23.5(28/119)
|
24.5(26/106)
|
0.001
|
a Fisher exact test were used.
Characteristics of live birth vs. non-live birth
Baseline characteristics, including female and male age, number of prior pregnancies and previous transfers, anti-müllerian hormone (AMH), and body massive index (BMI) were comparable for both groups. There were no statistical differences between the groups in terms of the number of embryos transferred, the protocols of endometrial preparation, days of embryo transfer and endometrial thickness. However, the serum β-hCG levels of patients with live births (median: 196 mIU/ml) was significantly higher than that of patients with non-live births (median: 140 mIU/ml, P=0.0001; Table 3).
Table 3 Characteristics of live birth vs. non-live birth
Characteristics
|
Live birth
|
Non-live birth
|
P
|
|
|
N
|
58
|
254
|
|
|
|
Female age (years)
|
31.9±4.7
|
32.5±4.9
|
0.346
|
|
|
Male age (years)
|
34.8±5.1
|
34.7±5.1
|
0.837
|
|
|
Infertility duration (years)
|
4.6±2.7
|
4.8±3.5
|
0.625
|
|
|
No. of previous gestation
|
|
|
0.242
|
|
|
%(n)
|
|
|
0
|
48.3(28)
|
42.2(107)
|
|
|
|
1-2
|
48.3(28)
|
47.6(121)
|
|
|
|
≥3
|
3.4(2)
|
10.2(26)
|
|
|
|
Causes of Infertility %(n)
|
|
|
0.805
|
|
|
Tubal/ Peritoneal
|
55.2(32)
|
48.0(122)
|
|
|
|
Ovulatory dysfunction
|
8.6(5)
|
10.2(26)
|
|
|
|
Male factor
|
10.3(6)
|
12.6(32)
|
|
|
|
Others
|
25.9(15)
|
29.1(74)
|
|
|
|
No. of previous transfer
|
1.8±1.0
|
2.1±1.3
|
0.187
|
|
|
AMH (ng/ml)
|
6.80±5.29
|
6.63±4.54
|
0.807
|
|
|
BMI (kg/m2)
|
21.9±2.9
|
22.3±3.5
|
0.434
|
|
|
Types of cycle
|
|
|
0.378
|
|
|
Natural cycle
|
22.4(13)
|
27.6(70)
|
|
|
|
Artificial cycle
|
77.6(45)
|
70.5(179)
|
|
|
|
Stimulation cycle
|
0.0(0)
|
2.0(5)
|
|
|
|
EMTa 5 days before transfer (mm)
|
8.6(8.0,10.0)
|
8.5(7.6,10.0)
|
0.327
|
|
|
Days of embryos transfer %(n)
|
|
|
0.807
|
|
|
5
|
58.6(34)
|
67.7(172)
|
|
|
|
6
|
36.2(21)
|
31.1(79)
|
|
|
|
both 5&6
|
5.2(3)
|
1.2(3)
|
|
|
|
No. of embryos transferred % (n)
|
|
|
0.574
|
|
|
1
|
39.7(23)
|
43.7(111)
|
|
|
|
2
|
60.3(35)
|
56.3(143)
|
|
|
|
|
|
Serum Pb
|
65.8±29.4
|
53.1±28.7
|
0.115
|
|
|
Serum E2b
|
1125(479,1386)
|
891(472,1871)
|
0.820
|
|
|
Serum β-hCG(mIU/ml)
|
196(144,221)
|
140(84,216)
|
2.40E-5
|
|
|
a EMT=endometrium thickness. b Only 68 patients were tested for serum progesterone levels and 62 were tested for serum estradiol.
Table 4 Effects of serum β-hCG levels on pregnancy outcomes by logistic regression analysis
|
|
95% Confidence Interval for OR
|
|
|
OR
|
lower limit
|
upper limit
|
P
|
Clinical pregnancy
|
1.875
|
1.522
|
2.310
|
1.38E-10
|
Early miscarriage
|
0.889
|
0 .739
|
1.069
|
0.889
|
Ectopic pregnancy
|
1.019
|
0.802
|
1.295
|
0.876
|
Live birth
|
1.416
|
1.162
|
1.726
|
0.001
|
Serum β-hCG was included as a categorical variable(categorized by 5-50, 51-100, 101-150, 151-200, 201-250, 251-299 mIU/ml).
Adjusted for female age, days of embryo transfer and number of embryo transfer.
Table 5 Thresholds of serum β-hCG level for prediction of clinical pregnancies and live births
Pregnancy outcomes
|
Clinical pregnancy
|
Live birth
|
|
Clinical pregnancy
|
Live birth
|
|
Threshold of β-hCG level(mIU/ml)
|
|
Threshold of β-hCG fold increase over 48h
|
|
58.8
|
108.6
|
|
1.4
|
1.9
|
AUC
|
0.752
|
0.649
|
|
0.899
|
0.808
|
95% CI of AUC
|
0.680-0.823
|
0.583-0.715
|
|
0.801-0.996
|
0.724-0.891
|
Sensitivity %
|
95.0
|
93.1
|
|
90.3
|
88.5
|
Specificity %
|
53.5
|
37.0
|
|
77.8
|
64.5
|
PPVa %
|
85.8
|
25.2
|
|
97.4
|
63.6
|
NPVb %
|
73.3
|
95.9
|
|
33.3
|
92.3
|
a PPV=positive predicted value; b NPV=negative predicted value.
Effect of serum β-hCG levels on pregnancy outcomes
Serum β-hCG levels had a positive effect on pregnancy outcomes (OR for a 50 mIU/ml interval), including clinical pregnancy (OR 1.875; 95% CI 1.522–2.310; P = 0.0001), and live births (OR 1.416; 95% CI 1.162–1.726; P = 0.001; Table 4).
Prediction of pregnancy outcomes
ROC analysis showed that the predicted value for clinical pregnancy was 58.8 mIU/ml with an AUC of 0.752 (95% CI: 0.680-0.823), a sensitivity of 95.0%, and a specificity of 53.5%. The threshold for live births was 108.6 mIU/ml with an AUC of 0.649 (95% CI: 0.0.583-0.715), a sensitivity of 93.1%, and a specificity of 37.0% (Fig. 1, Table 5). For the β-hCG fold increase over 48 hours, the cut-off for a clinical pregnancy was 1.4 with an AUC of 0.899 (95% CI :0.801-0.996), a sensitivity of 90.3%, and a specificity of 77.8%. The threshold for live births was 1.9 with an AUC of 0.808 (95% CI: 0.724-0.891), a sensitivity of 88.5%, and a specificity of 64.5% (Fig. 2, Table 5).