We retrospectively evaluated all patients having undergone hysteroscopic resection of a uterine septum at our department from 1997 to 2006. Primary outcome parameter was obstetric outcome (number of pregnancies, number of life births). As a second outcome parameter we evaluated the number and type of intra- and postoperative complications in order to gain information on the safety of the procedure.

A total of 64 women were enrolled in the study. Fourteen patients (22%) presented with primary infertility, defined as the inability to conceive after 12 months of contraceptive-free intercourse. Fifty patients (78%) presented with secondary infertility, defined as the inability to conceive after 12 months of contraceptive-free intercourse after having already conceived at least once. Twenty-two of 50 patients with secondary infertility (44%) have been diagnosed with recurrent abortion, defined as three or more consecutive spontaneous pregnancy losses before 20 weeks gestation. Written informed consent was obtained from patients for publication of the data presented. Copies of the written consents are available for review by the Editor-in-Chief of this journal.

All operations were performed as in-patient surgery. We neither used any form of long-term pre-treatment with danazol or gonadotropin releasing hormone (GnRH) agonists, nor pre-operative intravaginal administration of misoprostol in contrast to other investigators [¹⁵]. Uterine septoplasty was performed under general anaesthesia by two surgeons combining hysteroscopy and diagnostic laparoscopy. All procedures were performed in the follicular phase of the menstrual cycle, usually within seven days after the end of menstruation. In 38 women patency of the fallopian tubes was tested by chromo-pertubation with Lipiodol (Lipiodol Ultrafluid Ampul?, Guerbet, Austria) which was performed after septoplasty.

After cervical dilation with Hegar's dilators up to a width of 10 mm, an operative hysteroscope (Storz 10 mm fibreoptic resectoscope; Storz GmbH, Germany) was inserted. The uterine cavity was distended with a 1.5% glycine solution with a continuous irrigating flow. The hysteroscope's hook was placed in contact with the septum and an incision was made perpendicular to the septum. All operations were performed with monopolar electrosurgery, no scissors or lasers were used. The aim of the procedure was to accomplish a triangular and symmetric uterine cavity, which was achieved in all of the patients Uterine septoplasty was performed under laparoscopic supervision.

All patients were sent a questionnaire per mail. The following questions were evaluated: whether the patient tried to get pregnant after septoplasty; the number of pregnancies achieved; the number of miscarriages, extrauterine pregnancies, and spontaneous abortions, and the number of life births including information on gestational week and delivery mode; and whether patients had undergone any form of assisted reproductive techniques. All patients were asked to return the completed questionnaire and, in case that they had delivered in another hospital, to enclose a copy of the hospital delivery report or another document proving the life birth of a child.

Variables are described by frequencies and mean ? standard deviation (SD) of the mean. Differences between groups were analyzed by Chi square and Fisher's exact test. A p-value < 0.05 was considered statistically significant. Multiple comparisons were corrected by Bonferroni's correction. Statistical analysis was performed in SPSS 15.0.1 for Windows (SPSS Inc, 1989-2006).

We searched Medline and EMBASE and the Cochrane controlled trials register (search date: 20.12.2009; search terms: uterus subseptus, septate uterus, infertility, hysteroscopy, septoplasty) to identify cohort studies, systematic reviews, and meta-analyses evaluating the reproductive outcome after hysteroscopic septoplasty in women diagnosed with septate uterus. Studies were included if they were published as complete reports in English. Bibliographies of studies were searched for relevant citations. Multiple studies describing the same study population were included once. In this case, the original publication was used, i.e. the one with the earliest date of publication. Two authors assessed the eligibility of the studies and extracted data (KN and JO). Missing information and additional trials were not sought from authors. Data on pregnancy, life birth rates, complications, and re-hysteroscopy rates were extracted, pooled, and re-analyzed.

Fifty-seven of 64 patients (89%) underwent uterine septoplasty as described above including diagnostic laparoscopy and operative hysteroscopy. The mean age of the patients was 30.3 ? 5.7 years, the mean body mass index was 23.9 ? 2.7, and the mean duration of infertility was 4.5 ? 2.7 years. In 4/64 patients (6%) only operative hysteroscopy for dissection of the uterine septum was performed, whereas 3 patients (4%) underwent an additional curettage. Sixty-two patients had uneventful intraoperative and postoperative courses. However, in 2/64 (3%) patients, uterine perforation occurred. In both cases no adjacent organs were injured and the complication was managed by bipolar coagulation. No sutures were applied. No postoperative episode of fever was noted.

In 8/64 (13%) patients a second intervention was necessary since the uterine septum had not been completely removed during the first operation. The re-operation was done after a mean of 14.4 ? 10.7 months. Re-septoplasty was more often necessary in patients who had undergone hysteroscopy only than in those who had undergone combined hysteroscopy and laparoscopy (2/4 [50% vs. 6/45 [11%]; p = 0.07).

Complete follow-up was available for 49/64 (77%) patients. Mean follow-up time was 68.6 ? 25.2 months. All of the patients tried to become pregnant after septoplasty. The overall pregnancy rate after septoplasty was 69% (34/49). Eleven patients reported more than one pregnancy. Specifically, six women had two pregnancies, two women had three pregnancies, two women had four pregnancies, and one woman had six pregnancies, respectively. Thus, we registered a total of 49 pregnancies. Sixteen pregnancies had resulted from in vitro fertilization (33%), three pregnancies from insemination (6%), and the remaining 27 pregnancies (55%) resulted from spontaneous conception.

The overall life birth rate (LBR) was 49% (24/49). All life births were confirmed by hospital birth records. Three patients reported two life births, one patient reported three life births. Table 1 shows the reproductive outcome of women with and without a history of early spontaneous miscarriages and women with and without a history of recurrent miscarriages, ie three or more miscarriages. Women with a history of recurrent miscarriages had a significantly higher pregnancy rate after uterine septoplasty compared to women without a history of recurrent miscarriages.

The mean time interval between uterine septoplasty and the first delivery was 35.8 ? 22.5 months. Out of a total of 29 life births, 9 (31%) were preterm deliveries before 37 weeks gestation. No difference regarding preterm deliveries was observed comparing patients with or without a history of recurrent miscarriages (6/21 [29%] vs. 3/8 [38%], respectively; p = 0.6).

Figure 1 shows a graph of the time to first life birth in women after uterine septoplasty, demonstrating a slow, but consistent increase of life births during the observation period of 80 months. Cesarean section was the delivery mode in 12/29 (41%) women. The remaining 17 babies were delivered vaginally (59%). None of the patients was diagnosed with uterine rupture in the course of pregnancy and delivery.

Complete information on hormonal status, patency rates of the fallopian tubes, and the result of the partner's semen analysis was available for 21/64 (33%) women. Male factor, unilateral tubal factor, bilateral tubal factor, and hormonal pathologies, ie hyperprolactinemia, hyperandrogenemia, and thyroid dysfunction were noted in 14, 9, 3, and 11 of these women, respectively. Women with a bilateral tubal factor underwent in vitro fertilization. In 8/49 (16%) women, a change of the male partner was noted.

We identified 18 studies investigating the outcomes after hysteroscopic uterine septoplasty [^11-28]. One study was excluded from the analysis of obstetric outcomes, because data on pregnancy rates and LBR could not be extracted [¹⁷]. Table 2 provides an overview of the individual studies including our own data and a pooled analysis of 1501 women, yielding an overall pregnancy rate of 60.1% (892/1501) and a LBR of 45.0% (686/1501) after uterine septoplasty.

Details on postoperative complications and re-hysteroscopy rates are given in Table 3. For this analysis we excluded four studies [¹¹,¹²,¹⁶,¹⁸] with a total of 242 patients because the complication rates were not given in these studies. The pooled analysis of 14 studies with 1324 women undergoing uterine septoplasty resulted in an overall complication rate of 1.7% (23/1324). The overall rate of re-hysteroscopy was 6.0% (79/1324). There was no statistically significant difference regarding overall PR, LBR, and complications comparing our series with the pooled data from 17 studies in the literature (data not shown).

*after Bonferroni's correction

IUS = incomplete uterine septum; CUS = complete uterine septum; RM = recurrent miscarriages; PT = preterm delivery; PI = primary infertility; SI = secondary infertility; LM = late miscarriages

Ref = reference; n = number of patients; Year = year of publication; UM = uterine malformation; P = perforation; PF = postoperative fever; LSK = concomitant laparoscopy; PP = pelvic pain; CL = cervical laceration; RP = re-operation; POT = pre-operative therapy; IUS = incomplete uterine septum; CUS = complete uterine septum; GnRH = gonadotropin releasing hormone agonist; D = danazol; FOS = fluid overload syndrome; M = misoprostol in 2 patients 2 hours prior to surgery; PGE1 = prostaglandin E1; IU = intrauterine adhesions; PS = present study (Nouri et al. 2010); PA = pooled analysis