This retrospective study was conducted at Pondicherry Institute of Medical Sciences, a 650 bedded, tertiary care teaching hospital in south India. The study was approved by the Institute Ethics Committee. The case records of patients diagnosed with EP between October 2009 and September 2015 were retrieved from the medical records department. Patient characteristics like age, parity and risk factors for EP were noted. Mode of diagnosis, management modality, complications and need for blood transfusion were also recorded. The primary outcome measures studied were incidence of EP, their risk factors, mortality and morbidity in these women.

Data was entered in MS excel spreadsheet and analysed using SPSS software version 19.0. For categorical variables, data was compiled as frequency and percent. For continuous variables, data was calculated as Mean ± SD.

Over the six-year period, 7832 pregnancies were diagnosed, among whom 72 pregnancies were extra-uterine. The incidence of EP was 9.1/1000 pregnancies or one in 108 pregnancies. Majority of the women were aged 21-30 years [Table/Fig-1]. The most common risk factors were previous abortion (36.1%) and pelvic surgery (37.5%). Among the women who underwent pelvic surgery, 15 women had undergone tubectomy and two women had a tubal recanalization.

The classic triad of amenorrhoea, vaginal bleeding and lower abdominal pain was present in 29(40.3%) cases. Acute abdominal pain was the most common complaint, seen in 59 (81.9%) women; although a history of preceding amenorrhoea was present in 67 (93.1%) women. The other symptoms at presentation were vomiting and syncope. Cervical motion tenderness was elicited on pelvic examination in 58.3% cases.

A spot urine pregnancy test was performed in all cases and was found to be positive in 100% cases. A diagnosis of EP was made on clinical findings alone in 44(61.1%) women. However, ultrasonography was useful in making the diagnosis in 25 more cases. Ultrasonography was inconclusive in three cases, which needed a diagnostic laparoscopy to arrive at a diagnosis. The mean gestational age at diagnosis was 7.1 weeks. The site of ectopic was fallopian tubes in 68 cases (94.4%).

Patients who were haemodynamically stable, had a gestational sac size measuring less than 4cm by transvaginal ultrasonography, serum beta hCG (human Chorionic Gonadotropin) levels less than 10,000 U/ml and no free fluid in the pelvic cavity, were managed medically. Ability for regular follow-up was ensured before medical treatment. Fourteen (19.4%) women were managed medically with methotrexate. Nine of them had a single dose of methotrexate, while five needed a multiple dose regimen. Four of the 14 women, required surgery following failed medical management. Among the 62 women managed surgically, 47 were ruptured at the time of diagnosis, with haemoperitoneum seen intraoperatively. Mean haemoglobin at admission was 9.6 ±1.9 g/dL. More than half of the women needed blood transfusion (59.7%) and two women had transfusion related acute lung injury. Mean duration of hospital stay was 6.6±2.9 days. No deaths were noted. Admission to intensive care unit was required either due to haemodynamic instability or owing to complications like atelectasis and Transfusion Related Acute Lung Injury (TRALI). Abdominal wound infection was seen in four cases [Table/Fig-2].

EP accounts for 3.5-7.1% of maternal mortality in India [7,8]. The incidence of EP was 0.91% in our study. This is in agreement with most studies from developing countries where incidence ranged from 0.56-1.5% [1–3,5,6,9,10]. It is a significant cause of mortality in the first trimester. Timely referral to a higher centre is imperative in order to reduce mortality and morbidity.

The classical triad of abdominal pain, amenorrhoea and vaginal bleeding was seen in 40.3% of our cases. Other studies have reported this triad to be present in 28-95% women [1,11,12], clearly indicating that this is not a presenting feature in most cases. A history of amenorrhoea was present in 93.1% of our cases, but this may not come to light unless specifically enquired into.

Mean gestational age at diagnosis of EP was 7.1 weeks in our study, while Khaleeque et al., reported 6 weeks at diagnosis [2]. Singh et al., reported that 52% of their cases did not have preceding amenorrhoea [1]. Women may be unaware of an ongoing pregnancy and hence may not anticipate a pregnancy complication. Such women are most often seen first at a primary health centre or by a general practitioner and hence, the importance of careful history-taking cannot be overemphasised.

In our study, history revealed presence of at least one risk factor in 66.7% of the women, similar to other studies [1,2,13]. Among the risk factors, previous pelvic surgery was the most common (37.5%), followed by previous abortions (36.1%) in our study. Studies from various regions have reported a similarly high incidence of previous abortions, but contrary to our study previous abortions were the most common risk factor for EP in these studies [1–3,6,9,12,14,15]. The reason for previous pelvic surgery being the most common risk factor in our study could be attributed to the high caesarean section (33.6%) and tubal sterilisation (57.4%) rates in our state [16]. Singh et al., also reported prior tubal surgery as the most common (40%) risk factor in their study which is again a reflection of their high acceptance (57.4%) of tubal sterilisation as a method of family planning [1,16]. Hence a pregnancy test must be performed in all cases irrespective of their sterilisation status. Pregnancy must not be ruled out in women on such pretext.

History of Pelvic Inflammatory Disease (PID) was seen in 15.3% of our patients, similar to that reported by Singh et al., and Mufti et al., in their studies [1,15]. However, researchers from Nigeria have reported PID in 27-35.5% of their patients making it the most common risk factor for EP in that region. This high prevalence has been attributed to high polygamy rates in Nigeria [3,5,6].

Although clinical examination may raise suspicion of an EP, relying on clinical signs and symptoms alone would have missed the diagnosis in 38.9% of our cases. Ultrasonography was useful in diagnosing most of these cases, with the exception of three cases which needed laparoscopy for diagnosis. Culdocentesis, abdominal paracentesis and laparoscopy were used frequently in the past to aid diagnosis [3,5,6,17], but have been replaced by ultrasonography due to its non-invasive nature. Availability of point of care ultrasonography is of utmost importance in avoiding delay in diagnosis. This may not require gynaecologic specialists as physicians in community practice may be able to acquire comparable skills as residency graduates through a two-week intensive ultrasonography-dedicated training, as suggested by Jang et al., in their study [18].

Most of the cases (86.1%) were managed surgically and a salpingectomy was performed. A recent study concluded that laparoscopic surgical management was not better than a laparotomy in terms of the tubal patency and intrauterine pregnancy rates [19]. Most studies reported a similarly high rate of surgical management [2,3,6,9,20]. In contrast, surveys from the United Kingdom by Taheri et al., and van den Berg et al., reported a falling trend in the number of cases managed surgically (98% to 62 % and 50% to 27% respectively), over the last two decades [21,22]. This was attributed to the establishment of Early Pregnancy Assessment Units (EPAU) where EP is likely to be diagnosed at an early stage when medical management is still plausible.

The fallopian tubes were the most common seat of EP (94.4%). Most studies reported a higher incidence of EP in the right tube [1,3,6,10,23], while we found 72.2% of tubal ectopic pregnancies on the left side. In developing countries, majority of patients are diagnosed after tubal rupture. Our centre being a tertiary level referral centre, 65.3% of the women had ruptured ectopic pregnancies and presented with a haemoperitoneum, while some studies have reported 70-100% of ectopic pregnancies which were ruptured at diagnosis, mostly due to late referrals. Forty-three (59.7%)women needed blood transfusion, which was similar to that in other studies [2,9,10,20]. Udigwe et al., reported a 94.4% need for blood transfusion, as all women presented with a ruptured ectopic pregnancy and underwent a laparotomy and salpingectomy. Mean duration of hospital stay was 6.6 ±2.9 days. Udigwe et al., similarly reported that 94.4% of their patients had a hospital stay of less than 8 days, while 5.6% of the women needed prolonged hospitalisation upto 14 days [3].

There were no deaths due to EP during the period of our study. Maternal mortality due to EP was reported to be between 0% and 1.3% in various studies [2,3,5,6,9,20]. Mortality is mostly due to haemorrhage following rupture of the EP due to delayed referrals and diagnosis. National Institute for Health and Clinical Excellence has estimated that two-thirds of maternal deaths due to ectopic pregnancies in the UK may have been associated with inadequate care [24].

Prevention and treatment of PID and encouraging women to undergo an early transvaginal ultrasonography to confirm the location of pregnancy is likely to prevent late diagnosis. This will also allow medical management or fertility sparing conservative surgical management. Setting up Early Pregnancy Assessment Units has been shown to result in higher quality and cost-effective care, and to have a positive effect on early pregnancy care in the UK [22]. Future studies are required to evaluate usefulness of such EPAUs and feasibility of setting them up in India.

Ultrasonography being the mainstay for evaluating EP, its availability at the point of care will also help majority of patients by allowing safe and timely discharge of patients presenting to emergency departments with clinical suspicion of an EP [25]. Future research may be directed at assessing the impact of training doctors at primary and secondary levels of healthcare with two-week intensive ultrasonography courses, on the mortality and morbidity associated with EP.

Horne et al., suggested the use of serum Placental Growth Factor (PGF) to differentiate between intrauterine and ectopic pregnancies and concluded that serum PIGF was undetectable in women with tubal ectopic pregnancies and reduced, or undetectable, in miscarriage compared to viable intrauterine pregnancies [26]. Cabar et al., attempted to correlate the levels of serum Vascular Endothelial Growth Factor (VEGF) and ultrasonographic features in EP and concluded that serum VEGF was raised in ectopic pregnancies with cardiac activity [27]. Future studies are required to assess clinical utility of these markers.

Our study is limited by its retrospective nature. Further, we were unable to estimate the duration of delay in diagnosis and referral and its effect on morbidity.

Culdocentesis and laparoscopy have been superseded by non-invasive transvaginal ultrasonography and highly sensitive and accurate beta hCG assays for diagnosis of EP. Timely diagnosis and management in early pregnancy units with point of care ultrasonography can reduce the morbidity and mortality due to ectopic pregnancy. In conclusion, identifying underlying risk factors, availability of point of care ultrasound, complimented by availability of beta hCG assay and timely intervention will help reduce the morbidity and mortality due to EP.