Results
The first reported case of a duplicated uterus with obstructed hemivagina dates back to 1922, describing a sequential unilateral obstruction with hematocolpos, hematometra and hematosalpinx. However, no renal anomaly was mentioned [1] . In the 1970s, similar reports began to emerge. In 1971, Herlyn and Werner reported a case of a duplicated uterus and renal aplasia [2] . One year later, Amon et al .
[3] described a case of a bicornuate uterus with a double cervix and a double vagina associated with right hematocolpos, hematometra, hematosalpinx, partial atresia of the vagina and right renal aplasia. In 1976, Wunderlich also described a bicornuate uterus with a simple vagina and isolated right hematocervix without connection of the right uterus to the vagina in association with aplasia of the right kidney and ureter [4] . In light of these findings, Herlyn-Werner-Wunderlich syndrome (HWW) was described. This syndrome classically consists of a triad that includes a congenitally duplicated uterus, an oblique septum that generates a blind hemivagina and obstructs the menses outflow of 1 uterine cavity and an associated ipsilateral renal anomaly ( Fig. 2 ). However, the literature consistently reveals some heterogeneity in terms of anatomic Müllerian variants associated with this syndrome, which is likely the main reason why Smith and Laufer renamed the syndrome, only taking into consideration hemivagina obstruction and ipsilateral renal anomaly and suggesting the acronym OHVIRA [5] . Fig. 2 Classical Herlyn-Werner-Wunderlich syndrome. Fig 2
Classical Herlyn-Werner-Wunderlich syndrome.
The uterus, cervix, fallopian tubes and upper third of the vagina develop from the Müllerian ducts (also known as paramesonephric ducts) [6] . Müllerian congenital anomalies include developmental defects (agenesis/hypoplasia) and fusion defects (bicornuate uterus, didelphys uterus, septate uterus, transverse septum). Duplicated uterus is attributed to a failure in the fusion process of the Müllerian ducts at 9 weeks of gestation [7] . In a nonselected population, the prevalence of congenital uterine anomalies is approximately 5.5% [6] . Amongst the obstructed outflow forms that do not exhibit primary amenorrhea (ie, severe dysmenorrhea in the presence of menstrual cycles), HWW syndrome is the most common form [8] .
A comprehensive literature search of PubMed utilized the terms Herlyn-Werner-Wunderlich syndrome” or “obstructed hemivagina and ipsilateral renal anomaly” to review case reports and case series published prior to December 2020. The original articles were selected based on features such as age, major symptom, obstructed side, ipsilateral/contralateral renal anomaly, Müllerian anomaly, types of diagnostic imaging, obstruction mechanism and treatment. Data extraction compiled 521 into a descriptive analysis [ [12] , [13] , [14] , [19] , [20] , [21] , [22] , [23] , [24] , [25] , [26] , [27] , [28] , [29] , [30] , [31] , [32] , [33] , [34] , [35] , [36] , [37] , [38] , [39] , [40] , [41] , [42] , [43] , [44] , [45] , [46] , [47] , [48] , [49] , [50] , [51] , [52] , [53] , [54] , [55] , [56] , [57] , [58] , [59] , [60] , [61] , [62] , [63] , [64] , [65] , [66] , [67] , [68] , [69] , [70] , [71] , [72] , [73] , [74] , [75] , [76] , [77] , [78] , [79] , [80] , [81] , [82] , [83] , [84] , [85] , [86] , [87] , [88] , [89] , [90] , [91] , [92] , [93] , [94] , [95] , [96] , [97] , [98] , [99] , [100] , [101] , [102] , [103] , [104] , [105] , [106] , [107] , [108] , [109] , [110] , [111] , [112] , [113] , [114] , [115] , [116] , [117] , [118] , [119] , [120] , [121] , [122] , [123] , [124] , [125] , [126] , [127] , [128] , [129] , [130] , [131] , [132] , [133] , [134] , [135] , [136] , [137] , [138] , [139] , [140] , [141] , [142] , [143] , [144] , [145] , [146] , [147] , [148] , [149] , [150] , [151] , [152] , [153] ].
Data analysis was conducted on 521 cases reported in case reports and case series, which were organized into 2 groups. Group A (n=462) included cases diagnosed after birth, and Group B (n=59) included cases whose diagnostic investigation started prenatally.
Regarding age, Group A had a mean age at diagnosis of 16.64 years (range 0.5-65 years) when excluding case series that reported age with mean or median values.
In this group, the major symptoms/signs that prompted the diagnostic work-up where available, were abdominal/pelvic pain and cyclic pelvic pain ( Table 1 ). Table 1 Major symptom/sign a (Group A). Table 1 Abdominal/pelvic pain 142 (40,7%) Cyclic pelvic pain 128 (36,7%) Foul smelling vagina discharge 23 (66,0%) Abnormal vaginal bleeding 8 (2,3%) Infertility 8 (2,3%) Urinary retention 7 (2,0%) Other symptoms 33 (9,5%) a 113 cases not specified
Major symptom/sign a (Group A).
113 cases not specified
In Group B, as the syndrome was suspected prenatally, the main ultrasound features that prompted investigation were dysplastic/multicystic kidney (49.2%) and renal agenesis (39.0%). On rare occasions, the features reported were pelvic/abdominal cystic mass combined with renal agenesis or isolated pelvic/abdominal cystic mass ( Table 2 ). Table 2 Major symptom/sign (Group B). Table 2 Dysplastic/Multicystic kidney 29 (49,2%) Renal agenesis 23 (39,0%) Pelvic/abdominal cystic mass 3 (5,1%) Renal agenesis + Cystic mass 2 (3,4%) Dysplastic/Multicystic kidney + Cystic mass 1 (1,7%) Pelvic kidney suspicion 1 (1,7%)
Major symptom/sign (Group B).
There are multiple classification systems for Müllerian anomalies but there is no consensus on which system is more appropriate. The American Society for Reproductive Medicine (ARSM; formerly known as the American Fertility Society [AFS]) published a classification system in 1988, which has endured over time and is the most used system in the included articles. For that reason, this review uses the ARSM terms and definitions. A more recent classification system was published by the European Society of Human Reproduction and Embryology–European Society for Gynecological Endoscopy (ESHRE–ESGE 2013) for female genital tract anomalies, which classifies anomalies regarding the characteristics of the uterine body, cervix and vagina. When comparing both classification systems, uterus didelphys from the ARSM corresponds to bicorporeal uterus subtype b (U3b). In fact, when using the ESHRE-ESGE system, classic OHVIRA is classified as U3bC2V2. It is important to note that since 2021, ARSM adopted a more holistic classification system, taking into account the uterine body, cervix and vagina, similar to the ESHRE-ESGE system.
Regarding the type of uterine anomaly, uterus didelphys was the most commonly reported (88.7%), followed by bicornuate uterus and septate uterus ( Table 3 ). A diagram depicts all variants reported by type of Müllerian anomaly ( Fig. 3 , Fig. 4 , Fig. 5 ). However, the authors would like to highlight some limitations of the Müllerian anomaly classification. Historically, MRI was considered the gold-standard diagnostic imaging technique for assessing a Müllerian anomaly. Three-dimensional (3D) ultrasound imaging has improved and allows a very accurate characterization of Müllerian anomalies and often dismisses the need for further imaging techniques. The combination of an imaging technique with hysteroscopic and/or surgical findings, further increases the chance of correct classification of the anomaly. When reviewing all reported cases, we found that the uterine anomalies were classified by imaging techniques, such as ultrasound, CT scan or MRI, and by endoscopic/surgical findings. Sometimes, authors changed the imagiological classification because of surgical findings. This indicates that Müllerian anomalies classified by imaging techniques may not be precise, which suggest that using congenital anomaly classification as an outcome measure is questionable due to the lack of standardization of definitions and classifications. Table 3 Types of reported Mullerian anomalies according to the ARSM classification. Table 3 Uterus Didelphys 462 (88,7%) Bicornuate Uterus 34 (6,5%) Septate Uterus 25 (4,8%) Fig. 3 Reported variants in uterus didelphys. (A) Classical type with longitudinal obstructing vaginal septum (U3bC2V2); (B) Longitudinal non-obstructing vaginal septum (U3bC2V1); (C) Cervical aplasia (U3bC3V0); (D) Cervical communication (U3bC2V2 + Cervical communication) Fistula between obstructed hemivagina and nonobstructed cervix (U3bC2V2 + cervico-vaginal communication); (F) Cervical aplasia associated to a septum (U3bC3V2); (G) Classical type combined with fenestrated transverse septum (U3bC2V3 (nonobstructing septum)); (H) Septum microperfuration due to infection/distension (originally, U3bC2V2). Fig 3 Fig. 4 Reported variants in bicornuate uterus. (A) Bicornuate uterus bicollis with longitudinal obstructing vaginal septum (B) Bicornuate uterus unicollis with longitudinal obstructing vaginal septum; (C) Bicornuate uterus bicollis with septum microperfuration due to infection/distension. Fig 4 Fig. 5 Reported variants in septate uterus. (A) Complete septate uterus with longitudinal obstructing vaginal septum; (B) Partial septate uterus with longitudinal obstructing vaginal septum; (C) Complete uterus septate with cervical communication and longitudinal nonobstructing vaginal septum. Fig 5
Types of reported Mullerian anomalies according to the ARSM classification.
Reported variants in uterus didelphys. (A) Classical type with longitudinal obstructing vaginal septum (U3bC2V2); (B) Longitudinal non-obstructing vaginal septum (U3bC2V1); (C) Cervical aplasia (U3bC3V0); (D) Cervical communication (U3bC2V2 + Cervical communication) Fistula between obstructed hemivagina and nonobstructed cervix (U3bC2V2 + cervico-vaginal communication); (F) Cervical aplasia associated to a septum (U3bC3V2); (G) Classical type combined with fenestrated transverse septum (U3bC2V3 (nonobstructing septum)); (H) Septum microperfuration due to infection/distension (originally, U3bC2V2).
Reported variants in bicornuate uterus. (A) Bicornuate uterus bicollis with longitudinal obstructing vaginal septum (B) Bicornuate uterus unicollis with longitudinal obstructing vaginal septum; (C) Bicornuate uterus bicollis with septum microperfuration due to infection/distension.
Reported variants in septate uterus. (A) Complete septate uterus with longitudinal obstructing vaginal septum; (B) Partial septate uterus with longitudinal obstructing vaginal septum; (C) Complete uterus septate with cervical communication and longitudinal nonobstructing vaginal septum.
Due to the absence of ultrasound imaging features, it is not possible to attribute the septate/bicornuate uterus of ARSM (1988) a correct ESHRE-ESGE classification. All U3b (uterus didelphys) variants are reported in Table 4 and a schematic illustration is depicted in Fig. 3 . Table 4 Types of reported Mullerian anomalies according to the ESHRE-ESGE classification. Table 4 U3bC2V2 a 399 U3bC2V1 40 U3bC3V0 12 U3bC2V2 + Cervical communication 7 U3bC2V2 + Cervical-vaginal communication 1 U3bC3V2 2 U3bC2V3 (nonobstructing septum) 1 Total 462 b a 22 cases in which microperfuration has occurred due to infection/distension in a previous obstructing septum were regarded as longitudinal obstructing vaginal septum (V2) b Septate uterus [25] and Bicornuate uterus [34] were not included due to the lack of information regarding ultrasound features, which does not allow a feasible ESHRE-ESGE classification.
Types of reported Mullerian anomalies according to the ESHRE-ESGE classification.
22 cases in which microperfuration has occurred due to infection/distension in a previous obstructing septum were regarded as longitudinal obstructing vaginal septum (V2)
Septate uterus [25] and Bicornuate uterus [34] were not included due to the lack of information regarding ultrasound features, which does not allow a feasible ESHRE-ESGE classification.
One classification system suggests that the syndrome should be classified as type 1 or type 2 according to the obstruction mechanism (complete or incomplete) of the hemivagina. Type 1 represents complete obstruction, including type 1.1 (classic HWWS) and type 1.2 (cervical atresia). Type 2 represents incomplete obstruction, including type 2.1 with incomplete septum and type 2.2 with outflow through a connection to the contralateral cervix [156] . However, this classification does not include bicornuate and septate uteri. For obvious anatomic reasons, it is impossible to extrapolate this classification to these 2 Müllerian anomalies ( Figs. 4 and 5 ).
Regarding the type of anatomical obstruction, vaginal septum comprised 97.1% of all cases ( Table 5 ), and a schematic illustration is depicted in Fig. 2 , Fig. 3 , Fig. 4 . Table 5 Types of anatomical obstruction. Table 5 Longitudinal obstructing/nonobstructing vaginal septum 506 (97.1%) Cervical aplasia 12 (2.3%) Cervical aplasia + Longitudinal obstructing vaginal septum 2 (0.3%) Longitudinal obstructing vaginal septum (on the affected side) + Transverse nonobstructing vaginal septum (on the usually normal side) 1 (0.1%) Total 521
Types of anatomical obstruction.
Although the majority of OHVIRA diagnoses occur at an early reproductive age, usually a few months after menarche due to severe dysmenorrhea and pelvic pain, some patients have a later diagnosis. This subgroup of patients can be partially explained by incomplete outflow obstruction, even in the presence of a longitudinal obstructing vaginal septum. In this review, incomplete outflow tract obstruction was reported in 10.9% of all cases (58/521). The most common mechanisms were a longitudinal nonobstructing vaginal septum (75.8%) and cervical communication (15.5%), which allows menstrual blood flow between the 2 hemiuteruses. Rarer mechanisms include a combination of fenestrated septum associated with cervical communication, fistula formation and uterine cavity communication (as in partial septate uterus or partial bicornuate uterus) ( Table 6 ). Table 6 Anatomical location of incomplete obstructions. Table 6 Longitudinal nonobstructing vaginal septum 44 (75,8%) Cervical communication 9 (15,5%) Cervical-vaginal communication 1 (1,7%) Septum communication + Cervical communication 1 (1,7%) Uterine cavities communication 3 (5,2%) Total 58
Anatomical location of incomplete obstructions.
From a nephro-urological perspective, case series report a 63% rate of unilateral renal agenesis associated with uterus didelphys, which increase to 90% when associated with obstructed hemivagina [9] . As described, renal agenesis has been the classical form of urological pathology associated with the syndrome and has always been defined by radiological methods [10] . However, over time, some other forms have been reported, such as dysplastic, hypoplastic, multicystic, polycystic, atrophic and ectopic kidney [ 5 , 11 ]. The data in this review allowed the classification of 521 cases. Renal agenesis was by far the most common anomaly, with a rate of 88.5% ( Table 7 ), but the authors believe that this rate is overestimated. There are multiple reported cases of dysplastic kidneys with ectopic ureter insertion in the blind hemivagina, mainly in case series that are focused on the prenatal, neonatal and prepubertal periods [ 10 , 12 , 13 ]. In fact, there are some reported cases of dysplastic kidneys with ectopic ureters inserted into the hemivagina (with some residual function) that were not detected on imaging exams, including magnetic resonance [31] , in which a continuous urine leak was triggered after vaginal septum resection [ 17 , 18 ]. This has led to questions the true physiopathology of Herlyn-Werner-Wunderlich syndrome. Several authors have suggested that agenesis is the final stage of a spectrum, which results from a process of dysplasia secondary to the insertion of the ureter into a blind hemivagina, generating a major obstruction that compromises the normal development of the kidney. Invariably, this leads to the involution of the nephro-urological system ( Fig. 6 ) [ 10 , 13 , 14 ]. Right-sided vaginal obstruction and renal anomalies have been reported with higher frequency since the earliest reported cases [15] and a systematic review reported a 65% rate of right-sided anomalies [16] . In fact, of 512 reported anomalies (311 right-sided, 201 left-sided, 9 cases not included [8 not specified and 1 with contralateral renal anomaly]), we have verified a slightly higher rate for right-sided anomalies, supporting the literature (60.7% vs 39.3%). Of all analyzed cases, a very unusual case reported a renal anomaly associated with contralateral obstructed hemivagina ( Fig. 7 ) [16] . Table 7 Associated ipsilateral renal anomalies. Table 7 Agenesis 461 (88.5%) No specifications 454/461 Agenesis with ectopic ureter insertion to the vagina 7/461 Dysplasia 51 (9.8%) No specifications 9/51 Not specified ectopic/orthotopic kidney position with ectopic ureter insertion to the vagina 30/51 Ectopic kidney position 4/51 Ectopic kidney position with ectopic ureter insertion to the vagina 7/51 Orthotopic kidney position with ectopic ureter insertion to the vagina 1/51 Hypoplasia 3 (0.6%) Ectopic 1/3 Ectopic kidney position with ectopic ureter insertion to the vagina 2/3 Dysplasia and Hypoplasia a 3 (0.6%) No specifications 2/3 Ectopic 1/3 Dupplex collecting system 1 (0.2%) No anomaly, but contralateral renal anomaly ( Fig. 7 ) 1 (0.2%) a These 6 cases described as “hypoplastic” probably belong to the “dysplastic” group. Literature often mixes both terms due to similar/overlapping definitions. Fig. 6 Herlyn-Werner-Wunderlich Syndrome spectrum (A) Ectopic ureter insertion in the obstructed hemivagina; (B) Dysplastic process secondary to the ectopic ureter insertion in the obstructed hemivagina; (C) Renal involution generating agenesis with ectopic ureter insertion in the obstructed hemivagina. Fig 6 Fig. 7 Obstructed hemivagina with incomplete septum and contralateral renal anomaly. Fig 7
Associated ipsilateral renal anomalies.
These 6 cases described as “hypoplastic” probably belong to the “dysplastic” group. Literature often mixes both terms due to similar/overlapping definitions.
Herlyn-Werner-Wunderlich Syndrome spectrum (A) Ectopic ureter insertion in the obstructed hemivagina; (B) Dysplastic process secondary to the ectopic ureter insertion in the obstructed hemivagina; (C) Renal involution generating agenesis with ectopic ureter insertion in the obstructed hemivagina.
Obstructed hemivagina with incomplete septum and contralateral renal anomaly.
Symptoms usually start soon after menarche with severe dysmenorrhea due to partial outflow obstruction and, subsequently, a pelvic mass is found on clinical and/or imaging assessment. However, this syndrome may remain unrecognized for several years. One of the most likely causes is the existence of communication between the 2 hemivaginas [97] . In fact, in this review, when considering the presence of a communication, the median age at diagnosis was 19 years of age vs . a median of 14 years of age when completely obstructed. In addition, patient tolerance to pain and the use of painkillers may also may delay the diagnosis. The range of volumes from small volumes, such as 300-500 ml [ 75 , 85 , 86 , 88 , 96 , 110 ] to high volumes as 1000-2000 mL [ 54 , 111 ], mean that the distensibility degree of the vagina/uterus and the threshold to trigger pain are different for each person. On the other hand, with the advent of better ultrasound quality and imaging, prenatal diagnosis has been progressively added as a new presentation of renal agenesis or multicystic kidney, as an indirect sign of a Müllerian anomaly.
Even with severe dysmenorrhea, the presence of menstrual outflow delays the suspicion of an outflow obstruction. Menarche and physiological changes are easily evoked as justification for the symptoms. Introduction of a birth control pill relieves the menstrual symptoms, which further delays the true diagnosis. Due to the young age of the patients, most of the time, a full gynecological exam is not performed, which also interferes with a diagnosis. Therefore, a high suspicion index is needed to obtain the diagnosis.
Ultrasound imaging is an optimal screening exam for this condition and should be considered in cases of severe dysmenorrhea. Of the primary imaging diagnostic tools, ultrasound and magnetic resonance imaging (MRI) are the mainstays [ 53 , 154 ]. In this review, in 44.4% of cases, a combination of ultrasound and MRI was used as the imaging work-up. On the other hand, work-up was performed with ultrasound alone in 27.7% of cases. Other imaging exams, such as computed tomography scans [ 102 , 135 , 153 ], 3D-MRI [134] , 3D-CT scans [78] and colpography [105] have been reported.
A single-stage vaginoplasty with drainage of the obstructed side and resection of the septum is the gold standard treatment [ 5 , 81 , 139 ]. There are several reported surgical approaches, including, laparotomic, laparoscopic [125] and transvaginal approaches, which allow relief of the obstruction and the pain associated with the distension generated by retained menses. The transvaginal approach is the most commonly used and different techniques are reported, such as sharp dissection [22] , hysteroscopy with monopolar electrocautery [146] , hysteroscopy with bipolar electrocautery [40] , resectoscopy with bipolar electrocautery [46] , bipolar electrocautery [102] and CO 2 laser [136] . However, septum resection may not be possible when the vaginal septum is in a proximal position, and a hemihysterectomy may be necessary [139] . In fact, the greater the distance between the septum and perineum, the more likely a hemihysterectomy will be necessary [139] . However, due to the known variants, vaginoplasty may not be the best approach. The associated nephroureteric issues and complications related to long-term retrograde menstruation may need a different approach. A 3O (obstruction, ureteric orifice, and outcome) subclassification system associated with OHVIRA has been proposed to guide surgical management. This system considers the mechanism of obstruction, the presence/absence of ectopic ureter insertion and the possible long-term complications of the syndrome. In addition to vaginoplasty, other strategies, such as hemihysterectomy, cervicoplasty, ureteronephrectomy, salpingectomy and ovarian cystectomy, may be needed depending on the clinical scenario [155] .