Does lower gastrointestinal endoscopy during pregnancy pose a risk for mother and child? – a systematic review

An uncontrolled, retrospective, multicenter study in 1995 [10], which was published again one year later as an expanded cohort with added controls [8], reported 46 pregnant patients undergoing 48 sigmoidoscopies and 8 pregnant patients undergoing 8 colonoscopies. There were no differences in birth outcomes between the pregnant patients undergoing endoscopy during pregnancy compared to the pregnant patients not undergoing endoscopy during pregnancy. Both groups had similar indications for endoscopy. In addition, there were no differences in birth outcomes compared to the national American rates at that time. No adverse maternal events were reported following endoscopy. Following sigmoidoscopy, 4 voluntary abortions and 3 fetal demises occurred. All fetal demises were temporally and etiologically unrelated with the endoscopies. Following colonoscopy, there was one voluntary abortion and one fetal demise, both also temporally and etiologically unrelated with the colonoscopy.

In 2010, a study focusing exclusively on colonoscopies during pregnancy was published [3]. This retrospective, controlled cohort study reported on the safety and efficacy of colonoscopy in 20 pregnant patients. These pregnant patients were matched 1:1 with 20 pregnant controls with the same indication for colonoscopy but who did not undergo colonoscopy due to pregnancy. The study group was also compared to the pregnancy outcomes of the American national average. The majority of colonoscopies were performed in the second trimester of pregnancy (n = 16), with only 2 colonoscopies performed in respectively the first and third trimester. The study group trended towards worse pregnancy outcomes like stillbirth, premature delivery, low birth weight, low APGAR score, congenital defects and infant death after live birth, compared to the American national average. These non-significant differences can be attributed to the underlying illness in the study group according to the authors. When compared to the control group as described above, the study group tended to have slightly better fetal outcomes compared to the control group in terms of premature delivery, low birth weight, APGAR scores, congenital defects, neonatal ICU stay, infant postpartum hospitalization and infant death after live birth.

The 79 case reports describing 92 patients are summarized and categorized per trimester in Tables 1, 2 and 3.

Roughly, five major indications for endoscopy could be distinguished: (1) IBD and other colitis, (2) malignancy, (3) volvulus or incarcerated uterus, (4) non-malignant colonic obstruction and (5) gastrointestinal bleeding.

All temporally and etiologically related adverse events identified from the case reports are summarized in Table 4.

In the first trimester, 32 LGEs were performed in 30 patients. All complications following LGE in the first trimester are listed in Table 1. Three adverse events occurred within 1 week of the LGE. In one case report [20], the patient underwent sigmoidoscopy at gestational week 10 and the patient had an incomplete spontaneous abortion at 10.4 weeks. The patient suffered from severe rectal bleeding due to a heterotopic, abdominal pregnancy protruding the terminal ileum. This adverse event could possibly be attributed to the LGE, because this patient also underwent laparotomy after the LGE and suffered from severe gastrointestinal bleeding. The other two temporally related adverse events in the first trimester were both elective abortions, and were therefore classified as etiologically unrelated to the LGE [17,18].

In the second trimester, 39 endoscopies were performed in 35 patients. All complications following LGE in the second trimester are listed in Table 2. Six adverse events occurred within one week of LGE. Three cases reported three fetal deaths within one week of endoscopy. In the first case [51], the patient suffered from massive hematochezia due to multiple bleeding foci in the cecum and terminal ileum and underwent laparotomy shortly after colonoscopy. Fetal demise was evident several hours after surgery. This adverse event is possibly related to the LGE. The second patient was diagnosed with an advanced stage of colorectal carcinoma with liver metastases and ascites during pregnancy. After colonoscopy, the patient deteriorated rapidly and seven days after endoscopy fetal death was observed by ultrasonography. The mother died within 2 weeks after delivery [55]. This adverse event can probably be related to the LGE. The third case demonstrated a patient with progressive colonic distension caused by colonic pseudo-obstruction (Ogilvie’s syndrome). After colonoscopy, radiologic studies showed no evidence of colonic perforation, but the day after colonoscopy the abdominal distension progressed further, the patient went into spontaneous labor and the physicians decided to terminate the pregnancy [62]. This adverse event could also probably be related to the LGE. Two patients diagnosed with colorectal adenocarcinoma during pregnancy underwent elective abortion within one week of LGE in gestational week 16 and 20 [42,43] and in one patient labor was induced with prostaglandin in gestational week 26 [44]. These three adverse events were therefore classified as unlikely related to the LGE.

In the third trimester, 27 patients underwent 29 endoscopies. All complications following LGE in the third trimester are listed in Table 3. Four case reports demonstrated adverse events within one week of endoscopy. These four cases were likely related in one, possibly related in one and unlikely related in two of the cases. The first case describes a patient who was diagnosed with ulcerative colitis upon sigmoidoscopy in the sixth week of pregnancy. In the 28th week of pregnancy she exhibited signs of exacerbation and she underwent another sigmoidoscopy with biopsies. Following the second sigmoidoscopy, colonic perforation was suspected and an emergency caesarean section and exploratory laparotomy was performed. No colonic perforation was seen intraoperatively [11]. A live, healthy baby of 1054 g was delivered. This adverse event was classified as likely to be related to the LGE. The second patient was 33 weeks pregnant with twins, when she underwent two subsequent colonoscopies for the treatment and decompression of acute colonic pseudo-obstruction. She was already being treated with nifedipine upon presentation for inhibition of premature contractions, and nifedipine was stopped upon hospital admission. One day after the last colonoscopy at gestational week 34, she went into spontaneous labor and delivered healthy twins [85]. This adverse event is possibly related to the LGE. The third patient underwent sigmoidoscopy because of abdominal pain and distention in the 34th gestational week. Upon endoscopy, the splenic flexure appeared necrotic and the patient immediately underwent laparotomy with an emergency caesarean section [74]. This adverse event is unlikely related to the LGE. The fourth patient was diagnosed with a malignancy of unknown origin, and in the metastatic workup a colonoscopy was performed in gestational week 32. A poorly differentiated signet cell adenocarcinoma of the transverse colon was found, and after 4 days of dexamethasone administration for fetal lung maturation an elective caesarean section was performed [77]. This adverse event was unlikely related to the LGE.

One case report and one case series did not report at what gestational week the LGE was performed and were therefore not categorized. These case reports describe three pregnant women with IBD who underwent sigmoidoscopy for IBD disease assessment. One woman delivered a live baby of 1008 gram prematurely at 28.1 weeks [88]. A temporal relation was not found, and the authors do not link this adverse event to the sigmoidoscopy. In the case series, 2 out of 5 women underwent sigmoidoscopy, and one woman delivered a live baby prematurely. It is not reported if this woman underwent LGE [89].

A sensitivity analysis was performed by elongating the time span for the temporal relation between adverse events and the LGE. Initially, all adverse events were temporally related to the LGE if they occurred within one week after the LGE, however this analysis will classify all adverse events within three weeks of the LGE as temporally related. In the first trimester, this approach yielded no extra temporally related adverse events. In the second trimester, one additional temporally related adverse event was detected. In this case, the mother was diagnosed with advanced colorectal carcinoma during pregnancy and died together with the fetus two weeks after hospital admission around gestational week 23 [45]. This adverse event was unlikely to be related to the LGE. Finally, in the third trimester another seven temporally related adverse events were detected. Six premature deliveries were unlikely related to the LGE, as they were all elective caesarean sections [76,78-80,87] or induced labor [66]. The seventh patient suffered from ulcerative colitis and underwent LGE for assessment of disease activity in gestational week 32. Endoscopy showed the colon to be severely inflamed and two weeks later the patient delivered a premature baby of 1850 grams [64]. This adverse event is classified as probably related to the LGE.