INTRODUCTION

The World Health Organization (WHO) declared the coronavirus disease 2019 (COVID-19) a global pandemic in March 2020, having both direct and indirect impacts on UK healthcare provision. Emergency department attendance was reduced by 25% in the first week following government-imposed lockdown1, 2. The British Medical Association estimated that there were 1.32–1.50 million fewer elective hospital admissions in April, May and June 2020, with planned services in May 2020 operating at 31% of the preceding 2018–2019 average2, 3.

Both the Royal College of Obstetricians and Gynaecologists (RCOG) and the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) developed guidelines that aimed to rationalize early pregnancy and gynecological ultrasound service provision during the COVID-19 pandemic. Both advised that a scan should be carried out within 24 h for women with risk factors or symptoms associated with ectopic pregnancy (EP)4-6. In the event of an EP diagnosis, conservative management strategies were encouraged to minimize exposure to hospitals4. Surgery was recommended when no other management option was feasible4, 7. For all patients, a policy of telephone triage was introduced in the UK before women were able to access a hospital for the assessment of early-pregnancy problems.

There are concerns that a fear of attending hospital due to the risk of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection may have led women to delay accessing care for early pregnancy and acute gynecological problems. Findings from Italy and Israel suggest higher rates of surgery, ruptured EP and blood loss in women with EP during vs before the COVID-19 pandemic8-10. In New York, USA, 83% of women with EP presenting to an emergency department during the first wave of the pandemic were hemodynamically unstable, requiring urgent surgical intervention, compared with 23.5% in the previous year11. From an obstetric perspective, delayed presentation to the emergency department and subsequent increase in the number of emergency interventions secondary to the pandemic have been described12.

We aimed to describe and compare the characteristics of EPs presenting to an early-pregnancy assessment center in Inner London, UK, during vs before the COVID-19 pandemic.

METHODS Design and settings

This retrospective observational cohort study was performed at a dual center providing early-pregnancy and acute gynecological services in Northwest London, UK. All cases of EP for this study were identified on transvaginal sonography (TVS).

Relevant case records were identified via a search of the Astraia™ (Astraia Software GmbH, Ismaning, Germany) database over a 2-year period from 1 January 2019 to 31 December 2020, encompassing the time period prior to and during the COVID-19 pandemic, when the structure of our service was adapted according to national and international recommendations4-6. Search terms included a coded or non-coded diagnosis of EP or pregnancy outside the uterine cavity. Heterotopic pregnancies were included. Pregnancies were dated according to the last menstrual period or embryo transfer date.

Electronic case records were reviewed, and information related to predefined outcome measures was recorded as part of this retrospective analysis. The manuscript was written in accordance with the STROBE cohort study statement. The evaluated time period prior to the COVID-19 pandemic was between January 2019 and December 2019, and the time period during the COVID-19 pandemic was between January 2020 and December 2020.

In addition to delaying non-essential appointments, reducing follow-up appointments and reducing surgical management of miscarriage and EP, telephone triage was performed to rationalize service provision, in accordance with RCOG and ISUOG COVID-19 guidelines4-6. Measures were in place from January 2020 onwards, with the most stringent implementation between April 2020 and September 2020.

Outcome measures

Month of diagnosis, age, ethnicity, weight, mode of conception, gestational age at first TVS scan and at diagnosis, obstetric history, morphological and biochemical features of EP, admission and management were recorded.

Statistical analysis

Continuous data were assessed for normality using the D'Agostino–Pearson normality test. If data were distributed normally, a two-tailed t-test was performed, assuming both populations had similar SD. Normally distributed data are presented as mean with 95% CI. If data were not distributed normally, a two-tailed Mann–Whitney U-test was performed to compare ranks. Data with non-normal distribution are presented as median and interquartile range (IQR). Given the nature of the dataset and the study question, data were not paired.

Categorical data were analyzed using Fisher's exact test when comparing one or two variables and using chi-square test when comparing more than two variables. These data are presented as n (%). Statistical analysis was carried out using GraphPad Prism v8.2.1 (GraphPad Software, San Diego, CA, USA). A P-value < 0.05 indicated statistical significance.

Ethical approval

Ethical approval and written consent were not required, as this study involved retrospective review of case notes. As such, it was performed as a local clinical audit, as it was designed to produce information to inform on whether delivery of care was affected by the COVID-19 pandemic. Audit approval was obtained, local registration number GRM_071.

RESULTS

There were 21% fewer early-pregnancy and acute gynecological consultations during the COVID-19 pandemic compared to the previous year, with 12 657 consultations in 2019 compared to 10 026 in 2020 (Table 1). All consultations were face-to-face. The ratio of new (52%) to follow-up (48%) consultations was the same in both years (P = 0.88), with a lower proportion of blood tests performed in 2019 (2222 (18%)) vs 2020 (2184 (22%)) (P < 0.0001). There was no difference between 2019 and 2020 in the rates of miscarriage (18% vs 20%), pregnancy of unknown location (PUL) (7% vs 7%) or EP (1% vs 1%) observed in the unit (P = 0.38).

Table 1. Characteristics of 22 683 early-pregnancy consultations and diagnosed complications, before (2019) and during (2020) the COVID-19 pandemic Characteristic 2019 (n = 12 657) 2020 (n = 10 026) P Face-to-face consultation 12 657 (100) 10 026 (100) > 0.99* New consultation 6576 (52) 5219 (52) 0.88* Follow-up consultation 6081 (48) 4807 (48) Ratio of new to follow-up consultations 1.1:1 1.1:1 Blood test 2222 (18) 2184 (22) < 0.0001* Ratio of total consultations to blood tests 5.7:1 4.6:1 Complication 0.38† Miscarriage 2317 (18) 1996 (20) PUL 911 (7) 739 (7) Ectopic pregnancy 141 (1) 134 (1) Did not complete follow-up 13/141 (9) 9/134 (7) Data are given as n (%) or n/N (%), unless indicated otherwise. P-values were calculated using: * Fisher's exact test; or † chi-square test. PUL, pregnancy of unknown location.

A total of 275 EPs were diagnosed in the 2-year period, including 141 (51%) cases in 2019 and 134 (49%) cases in 2020 (Table S1). Thirteen (9%) EP patients did not complete follow-up in 2019, and nine (7%) did not complete follow-up in 2020 (Table 1).

Table 2. Demographic characteristics of 275 women diagnosed with ectopic pregnancy (EP), before (2019) and during (2020) the COVID-19 pandemic Characteristic 2019(n = 141) 2020 (n = 134) P Age (years) 32.2 [31.2–33.3] 33.0 [32.1–33.9] 0.25† Ethnicity 0.74‡ Asian 10 (7) 14 (10) Black 13 (9) 12 (9) Mixed 5 (4) 5 (4) White 34 (24) 50 (37) Missing data 79 (56) 53 (40) Weight (kg)* 68.0 (59.0–76.0) 67.0 (59.0–76.8) 0.74§ Conception 0.85¶ Natural 126 (89) 118 (88) Assisted 15 (11) 16 (12) PUL prior to EP diagnosis 52 (37) 56 (42) 0.46¶ EP diagnosis at first scan 89 (63) 78 (58) Data are given as mean [95% CI], n (%) or median (interquartile range). * Weight data were missing for 17 cases in 2019 and 28 cases in 2020. P-values were calculated using: † unpaired t-test; ‡ chi-square test; § Mann–Whitney U-test; or ¶ Fisher's exact test. PUL, pregnancy of unknown location.

Women with EP diagnosed prior to and during the pandemic were comparable in age (mean (95% CI), 32.2 (31.2–33.3) vs 33.0 (32.1–33.9) years; P = 0.25), ethnicity (P = 0.74) and weight (median (IQR), 68.0 (59.0–76.0) vs 67.0 (59.0–76.8) kg; P = 0.74) (Table 2). Method of conception was similar in 2019 vs 2020, with 126 (89%) vs 118 (88%) women conceiving naturally and 15 (11%) vs 16 (12%) women having assisted conception (P = 0.85). The majority of EP cases in both 2019 (89 (63%)) and 2020 (78 (58%)) were diagnosed on the first TVS examination, with the remainder having been classified initially as PUL in both groups (52 (37%) vs 56 (42%); P = 0.46).

History, sonographic and biochemical features

Thirteen (9%) patients in 2019 had a history of EP compared with 14 (10%) cases in 2020 (P = 0.84), with the majority having undergone surgical management (Table 3). The median (IQR) gestational age at the first TVS scan (41.0 (27.5–48.0) vs 38.0 (29.0–44.3) days; P = 0.13) and at diagnosis on TVS (42.0 (30.5–50.5) vs 41.0 (32.5–48.3) days; P = 0.34) were similar before and during the COVID-19 pandemic (Table 3 and Figure 1a,b). Both in 2019 and in 2020, most women had tubal EP (90% vs 93%; P = 0.67) (Table 3), with non-tubal cases including cornual, interstitial, myometrial, ovarian, cervical and Cesarean scar pregnancies.

Table 3. History, ultrasound and biochemical characteristics relating to ectopic pregnancy (EP) in 275 women diagnosed with EP, before (2019) and during (2020) the COVID-19 pandemic Characteristic 2019 (n = 141) 2020 (n = 134) P History of EP 13 (9) 14 (10) 0.84§ Missing data 0 (0) 2 (1) Two previous EPs 5/13 (4) 2/14 (1) Previous EP management Expectant 2 2 Medical 2 3 Surgical 12 11 Missing data 2 0 GA at first TVS (days)* 41.0 (27.5–48.0) 38.0 (29.0–44.3) 0.13¶ GA at diagnostic TVS (days)* 42.0 (30.5–50.5) 41.0 (32.5–48.3) 0.34¶ Tubal EP 127 (90) 124 (93) 0.67§ Missing data 1 (1) 0 (0) Mean EP diameter (mm)† 15.3 (11.0–22.3) 14.7 (11.5–18.3) 0.25¶ Largest EP diameter (mm)† 18.0 (13.0–24.1) 16.0 (13.0–22.0) 0.22¶ EP characteristics 0.62** Inhomogeneous mass 82 (58) 81 (60) GS only 36 (26) 30 (22) GS + YS 13 (9) 15 (11) GS + YS + live fetal pole 9 (6) 8 (6) Missing data 1 (1) 0 (0) Hemoperitoneum at diagnosis 33 (23) 35 (26) 0.58§ Missing data 0 (0) 2 (1) Deepest pool of hemoperitoneum (mm) 41.8 (31.1–51.4) 40.1 (28.5–54.5) 0.81¶ hCG at diagnosis of EP (IU/L)‡ 665 (224–2330) 1005 (412–2938)

0.03¶

hCG at diagnosis of tubal EP (IU/L)‡ 603 (182–2216) 952 (411–2544) 0.03¶ Data are given as n (%), n/N (%), n or median (interquartile range). * Gestational age (GA) data at transvaginal sonography (TVS) were missing for 12 cases in 2019 and eight cases in 2020. † EP diameter data were missing in two cases in 2019 and five cases in 2020. ‡ Human chorionic gonadotropin (hCG) level data were missing in two cases in 2019 and five cases in 2020. P-values were calculated using: § Fisher's exact test; ¶ Mann–Whitney U-test; or ** chi-square test. GS, gestational sac; YS, yolk sac.

Median gestational age (GA) at initial transvaginal sonography scan (a), GA at ectopic pregnancy (EP) diagnosis (b), and mean (c) and largest (d) EP diameters on transvaginal sonography, in women with EP, diagnosed before (2019) or during (2020) the COVID-19 pandemic. Whiskers are interquartile range.

EP size was documented by measuring the mean and largest diameters on TVS. There was no difference in the mean EP diameter (median (IQR), 15.3 (11.0–22.3) vs 14.7 (11.5–18.3) mm; P = 0.25) or largest EP diameter (median (IQR),18.0 (13.0–24.1) vs 16.0 (13.0–22.0) mm; P = 0.22) in the period prior to vs during the pandemic (Table 3 and Figure 1c,d). In 2019 vs 2020, similar proportions of women had an inhomogeneous mass (58% vs 60%), an empty gestational sac (26% vs 22%), a gestational sac containing only a yolk sac (9% vs 11%) and a gestational sac containing an embryo with cardiac activity (6% vs 6%) detected on ultrasound (P = 0.62). Hemoperitoneum was seen on TVS in 23% of cases in 2019 and in 26% of cases in 2020 (P = 0.58), with similar median (IQR) deepest pool measurements (41.8 (31.1–51.4) vs 40.1 (28.5–54.5) mm; P = 0.81) (Table 3).

For all cases with EP, median (IQR) serum human chorionic gonadotropin (hCG) levels were lower in 2019 compared to 2020 (665 (224–2330) vs 1005 (412–2938) IU/L; P = 0.03) (Table 3 and Figure 2). This difference in median (IQR) serum hCG levels between 2019 and 2020 was also observed when comparing only cases with tubal EP (603 (182–2216) vs 952 (411–2544) IU/L; P = 0.03) (Table 3).

Median serum human chorionic gonadotropin (hCG) levels at diagnosis in women with ectopic pregnancy, diagnosed before (2019) or during (2020) the COVID-19 pandemic. The asterisk indicates significant difference (P = 0.03). Whiskers are interquartile range.

Management and complications

Similar proportions of EPs were admitted for secondary care during vs before the COVID-19 pandemic (58% vs 63%; P = 0.46). There were significant differences between 2019 and 2020 when comparing the planned treatment strategy, with conservative management planned in 74 (52%) vs 85 (63%) cases and surgery planned in 66 (47%) vs 46 (34%) cases (P = 0.049) (Table 4).

Table 4. Secondary-care admission, management and biochemical characteristics and complications in 275 women diagnosed with ectopic pregnancy (EP), before (2019) and during (2020) the COVID-19 pandemic Characteristic 2019 (n = 141) 2020 (n = 134) P Admitted 89 (63) 78 (58) 0.46† Failed first-line management 8 (6) 21 (16) 0.01† Missing data 1 (1) 3 (2) Planned management 0.049† Conservative 74 (52) 85 (63) Surgical 66 (47) 46 (34) Missing data 1 (1) 3 (2) Final management 0.49‡ Expectant 39 (28) 32 (24) Medical 28 (20) 34 (25) Surgical 73 (52) 65 (49) Missing data 1 (1) 3 (2) hCG at diagnosis (IU/L)* Expectant management 219.0 (90.0–485.0) 312.0 (129.0–646.0) 0.32§ Medical management 815.5 (300.3–1582.0) 847.5 (465.8–1424.3) 0.54§ Surgical management 1571.0 (491.5–3310.0) 2664.0 (1076.5–6937.8) 0.01§ Ruptured EP 12 (9) 4 (3) 0.07† Missing data 0 (0) 3 (2) Data are given as n (%) or median (interquartile range). * Human chorionic gonadotropin (hCG) level data were missing in four cases in 2019 and four cases in 2020. P-values were calculated using: † Fisher's exact test; ‡ chi-square test; or § Mann–Whitney U-test.

However, the final type of EP management was similar prior to and during the pandemic, with the majority of patients undergoing surgical intervention (52% in 2019 and 49% in 2020; P = 0.49) (Table 4). Prior to the pandemic, 28% of EPs were managed expectantly, whilst 20% were managed medically with methotrexate. These rates were similar to those during the pandemic (P = 0.49), with 24% of EPs managed expectantly and 25% managed medically. Median hCG levels at the time of EP diagnosis in patients who were treated successfully using conservative measures were similar in 2019 and 2020 (expectant management: 219.0 vs 312.0 IU/L; P = 0.32, medical management: 815.5 vs 847.5 IU/L; P = 0.54) (Table 4).

More women required further treatment following failure of first-line management in 2020 compared with 2019 (16% vs 6%; P = 0.01). Median hCG levels at the time of EP diagnosis in women undergoing surgical intervention were lower in 2019 than in 2020 (1571.0 vs 2664.0 IU/L; P = 0.01). The rate of ruptured EP confirmed surgically was similar between 2019 and 2020 (9% vs 3%; P = 0.07) (Table 4).

SARS-CoV-2 positivity, self-isolation and delay in management

In 2020, 60% of women admitted for secondary care underwent SARS-CoV-2 polymerase chain reaction testing and obtained a negative result, while 40% were not tested (Table S2). Most women (88%) were not self-isolating at the time of admission; the remainder were admitted elsewhere or had missing details regarding self-isolation. No cases that were managed as outpatients or admitted locally experienced management delay due to the pandemic.

DISCUSSION

We found no difference in the mode of conception, gestational age at the first TVS examination and at diagnosis, location, size and morphology on ultrasound of EPs diagnosed in the year before vs during the COVID-19 pandemic. The rates of pregnancy complications and secondary-care admissions and the proportion of patients managed conservatively and surgically were also similar between the two cohorts. Among patients who underwent surgical management, the levels of serum hCG at diagnosis were higher during the pandemic compared to the year before the pandemic. Although the proportions of patients undergoing each management strategy were similar between the two years, we observed a reduction in the success of conservative management strategies during the COVID-19 pandemic. No woman with EP managed as an outpatient or admitted locally experienced management delay due to the pandemic. Although there were fewer patient visits in 2020 than in 2019, the number of new and follow-up face-to-face consultations, loss to follow-up and rates of early-pregnancy complications were similar.

The main strength of this study is the relatively large number of women with EP included, allowing appropriate detailed comparison of multiple outcome measures. The inclusion of patients from an Inner London center ensured that the cohort studied was diverse and the findings were more likely to be generalizable to a wider population. The identification of cases by electronic means reduced the possibility of any EP case not being included in the final analysis. This is reflected by the fact that, for many outcome measures, few data were missing.

The key weakness of the study was its retrospective nature, as the true impact of the COVID-19 pandemic may not be captured completely in the electronic notes. Another limitation is that this was a descriptive study.

We expected to find that women with EP presented later and with a higher rate of pregnancy complications, such as ruptured EP, during the pandemic. However, this hypothesis is not supported by the findings of this study. Our findings are also in contrast to those of the currently available literature on the impact of the COVID-19 pandemic on EP diagnosis and management, in which it has been suggested that hesitancy amongst patients in accessing healthcare resources led to an increase in the rates of surgery, ruptured EP, blood loss and hemodynamic instability8-13.

For women attending early-pregnancy assessment, a recent report suggested a 25% increase in the proportion of those with a diagnosis of miscarriage during the COVID-19 pandemic. This may reflect effective triage, with priority being given to those with complications14. Although virtual clinics were not implemented, telephone triage rationalized service provision, reducing patient visits during the pandemic. As all women with heavy bleeding or abdominal or pelvic pain were invited for prompt review, similar proportions of miscarriage, PUL and EP were reported4. However, even with rationalization measures in place, women were keen to avoid hospital settings1-6 due to the risk of contracting SARS-CoV-2 infection.

In the UK, the average ratio of new to follow-up consultations in an early pregnancy unit has been reported15 to be 1.88:1. Our prepandemic ratio of 1.1:1 reflects the greater number of follow-up consultations associated with having several second-opinion referrals, conservative miscarriage and EP management and use of a two-visit mathematical model for PUL. Perhaps, most relevant is our unit policy of allowing relatively liberal access to follow-up for patients, as we believe this may help ameliorate some of the known psychological sequelae of early-pregnancy complications16-22.

During the pandemic, although we reduced the number of visits in 2020, a considerable number of low-risk women would have had follow-up under normal circumstances. As higher-risk patients continued to attend, our new-to-follow-up consultations ratio remained relatively unchanged compared to that before the pandemic.

The de