1
applying various mathematical models), none of these methods showed the use of hCG to be useful
2
as a test for comprehensively and definitively diagnosing either an ectopic pregnancy or a viable
3
intrauterine pregnancy.
4
Consideration of clinical benefits and harms
5
The group recognised that some of the papers considered mathematical models. Whilst they
6
understood the intrinsic potential value that such models could have, the models considered in the
7
review had not been widely validated nor did they lend themselves to ease of use. Thus it was not felt
8
appropriate to recommend their use.
9
The evidence showed that the use of the rate of log change hCG or a change in hCG levels of greater
10
than 63% both provided a high negative predictive value but a low positive predictive value. In other
11
words, these measures were effective as tests to rule out the presence of an ectopic pregnancy but
12
not effective as tests to specifically identify the presence of an ectopic pregnancy. Thus, the group felt
13
that the value of the tests would be as a risk stratification tool to determine the urgency and type of
14
care that each woman requires, with the most urgent care being focused on those women in whom an
15
ectopic pregnancy was more likely.
16
The group recognised that whilst measuring the rate of log change hCG appeared to be a slightly
17
more accurate diagnostic test than simply measuring the rate of change of hCG, the findings were
18
similar for both approaches. It was felt that measuring an increase of greater than 63% would be
19
easier to calculate and a more useful measure in clinical practice than calculating the rate of log
20
change. In addition, the findings regarding the greater than 63% change were likely to be more valid
21
as the sample size of the study was much larger. As a result, the group agreed to recommend the use
22
of an hCG increase of greater than 63%. The evidence suggested that approximately 17% women
23
with a PUL would fall into this category and would have a high chance of having a viable intrauterine
24
pregnancy. Daus et al., (1989) reported that 60/357 women with a PUL had an hCG rise of greater
25
than 63%. Of these 54/60 were ultimately diagnosed with a viable IUP. Similar findings were reported
26
by Dart et al., (1999) who found that 52/307 of women with a PUL had an hCG rise of greater than
27
66% and 40/52 of these were ultimately diagnosed with a viable IUP.
28
The group agreed that for women with an hCG increase of > 63%, it would be appropriate to offer a
29
routine ultrasound scan within 7-14 days in order to confirm the location of the pregnancy.
30
The GDG interrogated the evidence to determine an appropriate lower threshold to identify those
31
women likely to have a failing pregnancy and for whom a different management strategy could be
32
used. Of the four papers which evaluated a decline of hCG, one (Daus et al., 1999) used ‘any
33
decline’, one used a decline of 36-47% (Morse et al., 2012), whilst the other two (Mol et al., 1998;
34
Dart et al., 1999) evaluated a decline of ‘greater than 50%’. These two studies suggested that 32-36%
35
of women with a PUL would fall into the latter category (63/195 and 109/307 respectively), and were
36
therefore likely to have a failing pregnancy. In these studies none of these women had a viable
37
intrauterine pregnancy and the risk of an ectopic pregnancy was low (about 1% when both studies
38
were combined). In Morse et al (2012), which used a threshold for decline of 36-47%, a higher
39
proportion of women (over 3%) were misclassified as having a miscarriage when they were eventually
40
diagnosed with an ectopic pregnancy. From these data and their own clinical experience, the group
41
felt that a decline in hCG levels of greater than 50% was highly likely to indicate a failing pregnancy
42
and that this was an appropriately conservative threshold to use.
43
The group agreed that for women with a decline in hCG levels of greater than 50% it would not be
44
necessary to conduct a repeat ultrasound scan, but instead that they should be asked to do a urine
45
pregnancy test in two weeks’ time. If this was negative and the woman was asymptomatic no further
46
action would be necessary. However, if it were positive then the woman should return to the
47
dedicated early pregnancy service for a further clinical review and individualised management by a
48
gynaecologist.
49
For the remaining group of women (i.e. those with an hCG level change between a decline of less
50
than or equal to 50% and a rise of less than or equal to 63%) it was agreed that review by a
51
gynaecologist would be warranted.
52
The group wished to highlight that hCG should not be used for a confirmatory diagnosis and that final
1
confirmation could only be provided by either an ultrasound scan, or a negative pregnancy test (to
2
identify a failed pregnancy).
3
The group considered the evidence that was available for the use of progesterone levels in
4
conjunction with hCG. They noted that for both diagnosis of ectopic pregnancy and diagnosis of viable
5
intrauterine pregnancy, there was little or no improvement in the negative predictive value of the tests
6
as compared with using hCG alone.
7
The group felt it important to stress the importance of symptoms over hCG. The group agreed that all
8
women, regardless of their hCG level, should be given written information about what to do if they
9
experience any new or worsening symptoms, including details about how to access emergency care.
10
Consideration of health benefits and resource uses
11
The GDG felt that the use of large numbers of hCG measurements was not an effective use of
12
resources, both in terms of the woman’s care and the cost of the tests. They wanted to avoid large
13
numbers of tests being performed without a diagnosis being made, and agreed that the use of more
14
than two hCG measurements should only be undertaken following review by a specialist health care
15
professional. Considering that there was no evidence of any added value of progesterone in making a
16
diagnosis and the performing the test has associated costs, the GDG agreed that progesterone
17
should not be used in the assessment of women with pain and bleeding and a PUL.
18
Quality of evidence
19
The quality of the evidence was either low or very low. The GDG was aware of the limitations of the
20
studies but ultimately agreed that there was sufficient evidence to make recommendations.
21
Information giving and psychological support
22
The GDG felt that it was important that women be given a realistic likely prognosis for their
23
pregnancy, based on their hCG levels, but also be informed that further confirmation would be
24
needed. For women whose pregnancies might be unlikely to continue, they thought that this would
25
avoid giving women false hope and ensure that they felt informed about the likely outcome. For these
26
women, the group also felt that it was very important that they be given information about where and
27
how to access support and counselling services.
28
Recommendations
29
Number
Recommendation
37 Assume women with a pregnancy of unknown location have an ectopic pregnancy
until the location is determined.
38 In women with a pregnancy of unknown location, place more importance on clinical symptoms than on human chorionic gonadotrophin (hCG) results, and review the woman’s condition if any of her symptoms change, regardless of previous results and assessments.
39 Use hCG measurements only for determining trophoblastic proliferation, to help decide ongoing monitoring.
40 Do not use hCG measurements to determine the location of the pregnancy.
41 Regardless of hCG levels, give women with a pregnancy of unknown location written information about what to do if they experience any new or worsening symptoms, including details about how to access emergency care 24 hours a day. Advise women to return if there are new symptoms or if existing symptoms worsen. 42 Take two serum hCG measurements 48 hours apart to determine subsequent
management of a pregnancy of unknown location. Take further measurement only after review by a senior healthcare professional.
43 Inform a woman with an increase in serum hCG concentration greater than 63% that she is likely to have a viable intrauaterine pregnancy. Offer her a tranvaginal