Talk:Paper - A case of early human ovarian pregnancy (1922)
A CASE OF EARLY HUMAN OVARIAN PREGNANCY
By JOHN I. HUNTER, M.B., Cu.M. Associate Professor of Anatomy, University of Sydney.
Tue present paper comprises a complete description of a specimen of primary ovarian pregnancy, of which I have already published a preliminary account ('21). The specimen was presented to me by Professor J. T. Wilson, who had received it from the Royal Prince Alfred Hospital, Sydney, diagnosed as a case of primary ovarian pregnancy. I desire to express my great indebtedness to Professor Wilson for the opportunity he has given me to describe this interesting specimen. My investigations confirm the original diagnosis. They also demonstrate, as the clinical records of the case indicate, that the pregnancy is at a very early stage of development. This fact renders the specimen of considerable value and justifies the present detailed description because it helps to throw light upon the process of imbedding and the phenomena occurring in the early stages of primary ovarian pregnancy which are still imperfectly known (cf. Ray ’21).
The specimen appears in the embryological collection of the Anatomy Department, Sydney University, under the catalogue number H 1538, and it will be referred to throughout this account under this designation.
HISTORY OF THE CASE
The patient was under the care of Mr Joseph Foreman, Honorary Gynaecologist of the Royal Prince Alfred Hospital, who directed that the ovary removed at operation should be examined for ovarian pregnancy.
The patient (G.R. aet. 37) on admission on 1.7.14, complained of pain in the hypogastrium of 12 hours’ duration. She had experienced a similar attack seven days previously in which the pain was accompanied by vomiting, rigor and perspiration. Her last menstrual period had appeared 22 days previously (9.6.14). Her periods were quite regular of the 28 day type, and consisted of a scanty flow of 8 days’ duration. The patient, who had been married five years, had two children; the first was four years of age, the second two and a half years. There had been no miscarriages. On 3.7.14 four days before the next period was due, an operation for ectopic pregnancy was performed. The abdomen was opened in the median line and clotted blood was found in the pelvic peritoneal cavity. The right ovary appeared to be a mass of blood clot, and was deemed necessary to remove. The tubes were apparently normal. The mesovarium was crushed by an angioclast and the free tissue - was removed with a scalpel.
The outstanding feature of this account is the regular menstrual history. Operation was performed four days before the time of commencement of the
Anatomy LYr 5 58 John I. Hunter
menstrual period which, in all probability, woyld have been missed. In this respect the history resembles that of the cases reported by Lea (’10) and Chiene (’18). In neither of the above-mentioned cases, however, were embryonic structures recovered and exact comparison with H 153 in which an embryonic papilla was found within an intact chorionic vesicle, is impossible. The chorionic vesicle of H 153 was greatly compressed. Its internal dimensions were 10 x 4-7 mm. and the villi varied from 2-4mm. in length. Histological examination of the sections obtained revealed a damaged embryonic papilla. The amniotic sac was found to be collapsed and ruptured. Blood from the intervillous space had apparently infiltrated into the embryonic papilla and invaded the yolk sac. The distorted papilla measures approximately 3-5 mm. in length (fig. 7).. No details of embryonic structure can be ascertained. The abdominal stalk and the wall of the yolk sac contain blood vessels within which nucleated red blood corpuscles may be seen. The general features indicate that the stage of development approximated to, but was probably earlier than, that of Graf von Spee’s embryo “Gle.”
H1538 greatly resembles the ovarian pregnancy described by Bryce, Teacher and: Kerr (08). In this case there was a history of seven weeks’ amenorrhoea. The chorionic vesicle measured approximately the same as H 153 and the villi were 2-3 mm. in length. There was a fragmentary papilla, the general characters of which resembled the “Gle” embryo. The dimensions indicate that the stage of development is approximately the same as in H 158.
It seems probable that H 153 is the result of fertilisation after the last menstrual period and if so the duration of gestation would be less than 20 days. However, as in Bryce, Teacher and Kerr’s specimen retardation of the growth of the vesicle and embryo has probably occurred on account of the abnormal environment, or differentiation may have ceased after the first attack of intra-peritoneal haemorrhage, so that the condition of the embryonic structures is not a true guide of the exact time of duration of gestation.
NAKED EYE DESCRIPTION OF THE SPECIMEN
The specimen removed at operation had been divided longitudinally and consequently both external and cut surfaces were presented for inspection. It consisted of an ovary (4 x 1-5 ems.) from the surface of which, in proximity to one pole, there projected a dark haemorrhagic mass of rounded contour - measuring 3 cms. supero-inferiorly and 2 ems. antero-posteriorly (figs. 1 and 2).
This mass rested upon the ovary for the distance of 2-5 cms. In its neighbourhood the ovary contained a corpus luteum of pregnancy measuring 14 x 8mm. The opposite pole of the ovary was occupied by a large atresic follicle. In one situation a mass of blood projected from the otherwise smooth surface of the haemorrhagic swelling; the smoothness of the surface of the swelling was due to the presence of an investing fibrous capsule which could be traced into continuity with the ovarian stroma. At the point of junction it attained its maximum thickness (1 mm.). It became extremely thin upon the A Case of Early Human Ovarian Pregnancy 59
surface of the haemorrhagic swelling and was apparently ruptured where the extravasation of blood was still represented by the projection of blood clot. The original cut had passed through the chorionic vesicle. In either
Fig. 1. Photograph of the ovary and the contained haemorrhagic swelling ( x 14).
Laminaled i gr ; Corpus Lureum Blood clo iz
Chorionic -- - alll Vesicle
Fig. 2. Diagrammatic representation of the specimen as shown in fig. 1.
portion of the specimen this structure could be seen lying excentrically in the haemorrhagic swelling. An examination of the blood clot in the vicinity of the chorionic vesicle revealed the presence of villi in section. Also several branching villi were seen lying free in the crumbled blood clot, The villi could Sen? 60 John I. Hunter
be traced towards the wall of the chorionic vesicle. The interior of the vesicle was comparatively smooth, but at this stage (infiltration with celloidin was already in progress) no embryonic papilla was detected.
The evidence so far obtained is sufficient to establish H 153 as a genuine case of primary ovarian pregnancy.
THE CRITERIA OF PRIMARY OVARIAN PREGNANCY WITH REFERENCE TO THE FULFILMENT OF THE REQUIREMENTS BY H153
The arrival of the definite diagnosis of primary ovarian pregnancy is greatly facilitated by the early stage of development of H 153. As Norris (’09) states, “it is much easier to make a positive diagnosis of primary ovarian pregnancy prior to the sixth or eighth week, than it is later after the gestation sac has grown to such proportions as to change its relation to the surrounding parts.” Freund and Thomé (’06), Webster (’04), Young and Rhea (711) and Caturani (714) have also stressed this fact.
It is necessary to exclude the following conditions which may closely simulate cases of primary ovarian pregnancy:
(a) Haematoma of the ovary due to a bleeding corpus luteum.
(b) Pregnancy in an accessory tube or in a diverticulum from the tube (cf. Webster, ’02).
(c) Implantation upon the fimbria ovarica; or upon an accessory fimbria which may become detached from the tube; or upon the fimbriated extremity of the tube.
(d) A ruptured tubai pregnancy in the intraligamentous position.
Until full and careful histological descriptions were available of early cases difficulty was experienced in ruling out the existence of all these conditions. In 1898, in his Ingleby Lectures, J. W. Taylor states that ‘‘early rupture of a tube from a pregnancy of two, five or six weeks standing is aspecial phenomenon of extra-uterine pregnancy which has as yet not received the recognition it deserves.” Now early operative treatment for ectopic pregnancy provides the best opportunities for securing such specimens as H 158, and in the absence of early stages the scepticism of Lawson Tait and Bland Sutton as to the existence of genuine cases of primary ovarian pregnancy was justifiable. Lawson Tait (’88) asserted that all so-called cases were primarily tubal. In an editorial upon the subject in 1900 the British Medical Journal declared that Webster, P. W. Taylor and Bland Sutton had adduced evidence which tended to confirm this opinion. Bland Sutton (’96) had stated that “ovarian dermoids and calcified foetuses from tubal ruptures into the broad ligament have been mistaken for primary ovarian pregnancy.” In discussing Croft’s fatal case (Croft, 700) in which a four months’ foetus was removed and in which the ovary could not be found on the side of the tumour Bland Sutton reiterated these views. He stated that Croft’s case was too far advanced to discuss dogmatically, but he would only be convinced of the possibility of the occurrence of A Case of Early Human Ovarian Pregnancy 61
primary ovarian pregnancy, if he were shown “an early embryo and its membranes contained in a sac within the ovary.”’ He was convinced of the genuineness of Van Tussenbroek’s classical specimen in 1901. After having made a special journey to Amsterdam to investigate the sections, he declared that he was convinced that this form of ectopic gestation was now possible (Bland Sutton, ’01).
In consequence of the difficulties of accurate diagnosis of primary ovarian pregnancy criteria have been laid down by various authors who regard the fulfilment of these conditions as essential to a positive diagnosis of ovarian pregnancy. German writers had accepted the possibility as early as 1850, and in 1878 Spiegelberg laid down his well-known criteria, viz.
(i) That the tube on the affected side must be intact.
(ii) That the foetal sac should occupy the position of the ovary.
(iii) That the sac should be connected to the uterus by the round ligament.
(iv) That ovarian tissue must be found in the wall of the sac.
Williams (’08) modified the last condition by stating that ovarian tissue must be found at several different sites in the wall.
Jacobson (’08) added two further requirements, viz.
(i) That an organic connection between the foetus and ovary must be demonstrated. This condition is sufficiently covered by the demand of Williams.
(ii) That the foetus should be visible in the cavity of the ovum.
The latter condition, which is reminiscent of the statements of Bland Sutton (700) and Mayo Robson (’02), is unwarranted, as Meyer and Wynne have pointed out (John Hop. Hosp. Bull. No. 838, 1919), because the development of the embryo is retarded by its abnormal environment and is often not recovered after rupture of the gestation sac. In Holland’s case (’11), in which the gestation had been of six weeks’ duration, the foetus had become converted into an amorphous mass. No embryo was found in 24 out of 48 cases collected by me from the literature. This series included only those cases usually regarded as positive and in which I could satisfy myself, from the descriptions at my disposal, of the presence or absence of the embryo.
Heincke (quoted by Caturani, ’14) adds the presence of placental tissue within the ovarian tissue as an additional criterion. But plasmodium may be formed independently of pregnancy, e.g. in cases of chorion-epithelioma of the ovary, sarcoma of the testis, and a dermoid cyst of the anterior mediastrium. Whitehouse (’10), who quotes the above examples, points out, however, that in them plasmodium did not occur alone. This is unlike his case, which he concludes is probably an example of ovarian pregnancy.
Norris (’09) emphasises the importance of examining the corresponding tube microscopically to exclude cases of tubal abortion. He states, “‘the tube should not only be intact, but should be microscopically free from any evidence of gestation.” Chiene (’18) disagrees with this statement and maintains that the removal of the corresponding tube as in his case may not be clinically essential or desirable. Lockyer (’17) nevertheless accepts Chiene’s case as 62 John I. Hunter
a genuine example of primary ovarian pregnancy. Further Caturani (’14) maintains that a decidual reaction may be present in the corresponding tube - in cases of primary ovarian pregnancy.
Of the requirements detailed above, those fulfilled by H 153 are as follows:
(i) The tube on the affected side was intact (macroscopically).
(ii) The tumour is part of the ovarian mass which was connected to the uterus by the ligament of the ovary.
(iii) The ovarian stroma encapsulates the chorionic vesicle or blood clot.
(iv) Chorionic villi are abundantly present.
(v) An embryonic rudiment is visible within the cavity of the ovum.
The condition of Norris that the corresponding tube should be examined microscopically is unfulfilled. However, tubal abortion is not difficult to exclude. The ovarian stroma completely encapsulates the haemorrhagic swelling which contained the chorionic vesicle, and the ovary was quite free from surrounding structures so that at operation the specimen could be removed by the application of an angioclast to the mesovarium.
In this connection it may be pointed out that in van Tussenbroek’s classical case the tube was not subjected to microscopic examination. Bryce, Teacher, and Kerr (’08) state in reference to this specimen that “the Fallopian tube was in no way attached to the ovary and, being normal, was obviously not the primary seat of implantation of the ovum.” In an editorial of the British Medical Journal (1900, p. 922) the statement is made that van Tussenbroek’s specimen proves that ovarian impregnation is a fact no longer to be denied. The existence of a considerable mesovarium and the absence of adhesions are facts stressed in this summary. These desiderata are present in the case of H 153.
DESCRIPTION OF THE SECTIONS
The following description is based upon incomplete celloidin sections 10 u in thickness and complete celloidin sections 40 to 50 in thickness. The thick sections proved of considerable value in presenting the topographical relationship of tissues upon which certain conclusions which follow are partly based.
The ovarian capsule.
The capsule is everywhere composed of fibrous tissue. In many parts the nuclei are sparse and the stroma cells can be seen to be arranged in concentric laminae around the blood clot similar to the specimen described by Hewetson and Jordan-Lloyd (’06). In addition there are considerable patches of connective tissue in which the nuclei are abundant. Both types of connective tissue abound throughout the ovary. As already revealed by nakedeye examination the capsule of ovarian tissue was ruptured for the extent of 9mm. from which blood clot projected. The capsule was fragmentary for some distance beyond the site of rupture. The superficial aspect of this part of the capsule was covered by a thin layer of blood. It is evident that the A Case of Early Human Ovarian Pregnancy 63
rupture of the capsule caused the intraperitoneal haemorrhage found at operation. Such a rupture is probably brought about by a considerable increase of blood surrounding the chorionic vesicle so thinning the capsule until it finally gives way under the internal pressure. ‘
Lea (710) has pointed out that intraperitoneal haemorrhage in cases of primary ovarian pregnancy may be produced by the penetration of the ovarian capsule by syncytial elements, especially when the implantation cavity is in a superficial position. This occurs relatively infrequently. Examples of its occurrence are provided by the cases described by Boesebeek (’04), Norris and Mitchell (’08), Banks (’12), and Freund and Thomé (’06).
Fhe chorionic villi.
The chorionic villi are to be seen on all aspects of the compressed chorionic vesicle although they are especially abundant at one pole adjacent to the ovary. Here the intervillous spaces are intact. A similar distribution occurred in the cases described by van Tussenbroek (’99), Hewetson and Jordan-Lloyd (706), and Bryce, Teacher, and Kerr (’08).
Fig. 3. Chorionic villus and wall of the chorionic vesicle showing embryonic blood vessels ( x75). 64. John I. Hunter
The villi consist of a loose connective tissue core composed of elongated spindle-shaped cells with well-defined oval nuclei. This connective tissue is continuous with the extra embryonic mesoderm composing the wall of the chorionic vesicle (fig. 3). Blood vessels containing nucleated red blood corpuscles occur in both these situations. The usual double covering of the connective tissue core is present. The cyto-trophoblast consists of a distinct layer of cubical cells—Langhan’s layer. The outer layer consists of plasmoditrophoblast in which nuclei staining darkly with carm-alum are often to be seen arranged in successive rows. Irregular plasmodial processes are also present, some of which abut against the extravasated blood (fig. 4).
The relationship of the villi to the ovarian tissue.
It has already been pointed out that in some cases syncytial elements perforate the capsule of ovarian tissue and so produce intraperitoneal haemorrhage. In Freund and Thome’s case (’06) the perforated capsule was
Fig. 4. Showing villi on the external aspect of the chorionic vesicle imbedded in fibrin which separates them from the capsule of ovarian stroma ( x 35). (D. J. Farrell, del.)
1-2 em. in thickness. In the cases of Thompson (’02), Young and Rhea (11),
Hannes (712), and Caturani (14), the villi had invaded the surrounding stroma
without causing perforation. H 153 is, on the other hand, to be included
with the specimens described by Norris (’09), Mall and Cullen (’13), Kelly and A Case of Early Human Ovarian Pregnancy 65
McIlroy (06), Leon and Holleman (’02), Hewetson and Jordan-Lloyd (‘06), Chiene (’18) and Lea (’10), in which the villi do not invade the stroma. A laminated blood clot separates the villi from the capsule of ovarian tissue. Weigart’s fibrin stain reveals in H 153 masses of fibrin surrounding the villi, while a firmer laminated clot abuts against the capsule (fig. 4). Holland (711) ascribes such a condition in the case described by him to recurrent haemorrhages. Recurrent haemorrhages with a very slow circulation of blood in the implantation cavity would account for the appearance in H 153 also. This blood clot in most situations closely surrounds the villi. In one position,
Cavity of Chorionic Vesicle |,
Deep layer of stroma (detached by blood
effusion)” a ot
Ovarian _ Stroma
Blood clot (surface effusion )
Laminated Blood Clot . (infiltration within the stroma) Capsule of Ovarian Tissue
Fig. 5. Showing the intervillous spaces which are preserved at one pole of the chorionic vesicle. Between the surface and the spaces are to be seen three layers, viz. the capsule of ovarian tissue, blood extravasation, and a zone of barrier decidua-like cells ( x 20). (D. J. Farrell, del.)
viz. where the villi were found to be most abundant, intervillous spaces are still preserved (fig. 5).
These spaces are bounded by a zone of cells which walls off the laminated blood clot. I believe that this layer originally belonged to the capsule of ovarian tissue. Separation has apparently been effected by an infiltration of blood into the ovarian capsule. Consequently, where the intervillous spaces are present, they will be bounded by a thin necrotic zone of tissue. This zone consists of decidua-like cells. It is separated by a blood-filled interval, 2 to 5 mm. in width from the main mass of ovarian capsule. Therefore, three zones 66 John I. Hunter
intervene between the intervillous spaces and the surface of the haemorrhagic swelling, viz. the ovarian capsule, a laminated blood clot, and a barrier zone of decidua-like cells (fig. 5).
An important aspect of this condition is that the deep layer of stroma is very thin, It is on this account that the villi do not establish a firm connection with the stroma, for, in effect, their original connections have been torn away by the blood extravasated into the stroma which encapsulates the implantation cavity. This haemorrhage has distended and eventually ruptured the capsule in H 153, so producing the intraperitoneal haemorrhage which led to operative interference. In the specimen described by Bryce, Teacher, and Kerr (’08) a similar separation of the villi from the capsule has been effected, although in one situation villi came into direct relation with the stroma of the ovary. The authors state that “‘it is apparent from a study of the sections that considerable haemorrhage has taken place shortly before the operation, and that the blood has more extensively infiltrated the tissue intervening between a layer of necrotic tissue immediately applied to the villi and the more healthy living ovarian tissue forming the outer lamella of the wall of the cavity. A large coagulum was thus formed round the chorionic vesicle and the whole constituted a “fleshy mole” which would doubtless, in short time, have been extruded from the ovary into the peritoneal cavity.” This describes accurately the condition found in H 158. Webster (’07) stated that haemorrhage had occurred in the ovarian stroma in his specimen of ovarian pregnancy especially near the placental site. In the Bryce-Teacher uterine ovum a similar haemorrhage was found in the decidua but no extensive infiltration had ensued (cf. Bryce and Teacher (08), Pl. III).
The marked disturbances of the circulatory mechanism in cases of ovarian pregnancy account, I believe, for the frequency in which the embryo is degenerated or absent. As Kelly and MclIlroy (’06) point out the embryo is likely to perish at an early age although the envelopes go on living. Consequently retardation of development of the embryo is not infrequent (cf. Holland (’11), van Tussenbroek (’99), Bryce, Teacher, and Kerr (’08)). Freund and Thomé (706) believe that the embryo develops normally or not according to the size of the vessels in the region of the implantation cavity and regard the vicinity of the hilum as being the most favourable site. But, on the other hand, a profuse extravasation of blood into the ovarian stroma in the manner shown to have occurred in H 158 would clearly interfere with normal intervillous blood flow and so retard development. Where the intervillous spaces are not preserved degenerating villi are to be found, while the fewness of villi in these situations indicates that some have already been absorbed. In the ovary, as in the tube, lack of decidual reaction has been cited as the cause of the abnormal circulatory condition in the vicinity of the ovum. In a specimen so young as H 158 it is a matter of great interest to determine if there is any definite decidual reaction to be found. A Case of Early Human Ovarian Pregnancy 67
THE REACTION OF THE MATERNAL TISSUES
In H 158 there is a general increase in vascularity of the ovarian stroma which is highly cellular in character. There are hosts of leucocytes in the blood clot against which the villi abut. A similar infiltration of leucocytes has been observed by Mall and Cullen (’13) and Hewetson and Jordan-Lloyd (’06). In general the condition found in H 153 agrees with the statement of Kelly and McIlroy (06) that the reaction of the ovary in ovarian impregnation resembles that of the uterus in uterine pregnancy. But these authors conclude their discussion of this subject by stating that ‘“‘nowhere throughout the specimen can tissue be found which has taken upon itself the function or appearance of a decidual layer.” They state that lutein cells are present in the fibrin layer which together with some connective tissue cells separate the foetal tissues from the corpus luteum. These decidua-like cells, they believe, are the cells which have been mistaken for true decidual cells by some writers. van Tussenbroek (’99) found large granular necrotic cells which she first interpreted as being due to a decidual reaction. Later she was convinced that these cells were lutein and connective tissue cells. Banks (’12) and Graham (’12) interpret the decidua-like cells present in their specimens as. being at Jeast partly trophoblastic in origin.
Only a small minority of observers admit that the decidua-like cells may be the result of a maternal reaction. Bryce, Teacher, and Kerr (08) acknowledge that many of the large mononuclear cells spread out in the ovarian stroma in their specimen may be maternal in origin. Whitehouse (’10) found similar cells. Caturani (’14) described an attempt at decidual reaction especially around the blood vessels of the stroma. Webster (’04) found a definite decidual reaction lining the intervillous spaces. He denies, however, that ovarian tissue itself is responsible for this change and attributes the reaction to embryonal inclusion of Mullerian tissue in the ovary. In Giles and Lockyer’s case (14) the ovarian stroma contained many large blood vessels and showed decidual reaction in the medulla and cortex. Norris (’09) states that in the specimen described by Franz (’02) there were a few small groups of decidual cells immediately around the ovum. Norris concluded, however, that a study of the literature proved that few authors record “the presence of decidua-like cells and in these the identification is somewhat doubtful.” Williams (’08) who also summarised the literature upon this subject was also convinced that a definite decidua was absent in cases of ovarian pregnancy.
The survey of the literature given above, illustrating as it does that the majority of observers speak against a decidual reaction, would tend to convey the impression that the conclusion arrived at by Williams and Norris is a correct one.
After an investigation of H 153 with reference to the occurrence of a true decidual reaction, I am convinced that this condition does occur in the early stages of ovarian pregnancy, so confirming the conclusion of Giles and Lockyer 68 John I. Hunter
(714). In discussing the relationship of the chorionic villi to the maternal tissues, the appearances were interpreted by supposing that blood had infiltrated into the ovarian stroma, so separating a thin layer of cells from the main mass and tearing this and the villi away from their connexions with the capsule of connective tissue. This conclusion is based upon the presence of a definite
INFILTRATING oe THE STROMA
10,00 0°o%G' 3
Fig. 6. The barrier cell zone is shown separated from the ovarian stroma by means of an extravasation of blood ( x 60). (D. J. Farrell, del.)
layer of cells which is found to bound the maternal aspect of the intervillous spaces (fig. 6). This layer is most perfectly seen where the intervillous spaces are best preserved. If it be followed, it is found to form an imperfect barrier of cells. This is interrupted at various points on account of the haemorrhage in the stroma being in many situations continuous with the blood in the A Case of Early Human Ovarian Pregnancy 69
intervillous spaces. There is no interruption of the barrier layer in one position for the extent of 8 mm. and the layer can be followed on the opposite side of the blood clot which has broken through it (fig. 5). On the external aspect of the chorionic vesicle, it is only represented here and there by groups of cells, which still show the barrier arrangement. But here the intervillous space is imperfect, being continuous to a considerable extent with the blood within the stroma. This is the site at which the villi have already been noticed in a state of degeneration. .
The barrier is made up of large elongated, polygonal, oval and pearshaped cells. These are mainly placed concentrically around the intervillous space. The cytoplasm of the cells is stained uniformly with carm-alum and the nuclei are large and conspicuous. With haematoxylin the nuclei are again deeply stained in some of the cells. In others the nucleus is large, oval and. more faintly stained. 7
The decidua-like cells are imbedded in a fibrin layer, scattered cells extending outwards towards the ovarian stroma. In a few sections the ovarian. stroma surrounding the implantation cavity has not been infiltrated with blood, and. consequently the barrier zone and ovarian stroma are, in these sections, found to be continuous with one another for the extent of 3mm. Here there is a definite decidual reaction of the stroma and transition stages are to be seen between large darkly staining cells bounding the intervillous space and fusiform connective tissue cells. Similar transition forms were noticed in Hewetson and Jordan-Lloyd’s case (’06). This indicates the origin of the large cells from the ovarian stroma. These cells correspond in arrangement, shape and staining reactions with the cells of the barrier layer which ~ lines the intervillous spaces elsewhere. The tissue formed by these cells though not extensive represents an ovarian decidua.
I believe that H 153 illustrates a retrogressive phase with regard to the development of the decidua. The infiltration of blood into the stroma which has separated the barrier zone from the capsule proper has also led to the partial disintegration of this deep zone. Consequently in some situations it is only represented by isolated groups of cells. Extreme difficulty would have been encountered in distinguishing these groups from trophoblastic masses if the continuous lamina of cells forming the ovarian decidua in H 1538 had been wholly destroyed.
The difficulty of distinguishing these two types of cells from one another would also have been increased because the stroma of the capsule in H 158, where it has been isolated from the villi in the intervillous space by a considerable blood filled interval, does not present the intermediate stages in the formation of the large decidual cells. These facts probably explain why the decidual reaction in other specimens is less marked than in H 153. It may also explain why isolated decidua-like cells have been found in many specimens which have been interpreted as being wholly foetal in origin. The truth is . that the uterus is imitated by the ovary in its reaction not only by leucocytic 70 John I. Hunter
infiltration, by increased vascularity and by increased cell formation, but by the formation of true decidual tissue. However, the badly regulated blood flow in the ovary brought about when the plasmodium invades the vessels, leads, in many cases, to bleeding beyond the intervillous space into the ovarian stroma. If this is not arrested, the inner layer of decidua cells becomes stripped away from the stroma and destroyed, and the enlarging cells in the latter resume their normal size, being no longer subject to the direct influence of the trophoblast. As already shown, the uterine decidua surrounding the Bryce-Teacher ovum exhibits a similar infiltration, but this is not extensive, though its effect in this case also was to strip off a zone of maternal tissue. In the ovarian pregnancy of Bryce, Teacher, and Kerr (’08), the infiltration was of a much greater extent than in the uterine decidua, and this fact probably accounted for the relatively few decidua cells found.
Caturani (14), in speaking of van Tussenbroek’s specimen, believes that the lack of decidual reaction accounts for the considerable haemorrhage between the maternal tissue and the oval sac. But I have presented another aspect of the case, viz. that the invasion of comparatively large vessels imbedded in fibrous tissue will probably lead in cases of ovarian pregnancy to relatively profuse haemorrhage independently of the degree of decidual reaction at this stage. I have endeavoured to show that this haemorrhage ploughs up the decidual tissue and separates the villi from their close connexion with the ovarian stroma and consequently the degree of decidual formation is greatly reduced.
A priori, there is no reason why the ovarian stroma cannot react in a manner similar to that of the endometrium. Indeed, decidual reaction has been found on the surface of the ovary in cases of uterine pregnancy as part of the condition called by Taussig (06) “ectopic decidua formation.”
It has been noted above that some observers, e.g. van Tussenbroek (’99), regard decidua-like cells described by them as being lutein in nature. The following description of the corpus luteum and of the relationship of the implantation cavity to it, makes this explanation untenable in the present case.
THE RELATIONSHIP OF THE CHORIONIC VESICLE TO THE CORPUS LUTEUM
A corpus luteum corresponding to the duration of gestation as indicated by the clinical history and confirmed by the degree of development of the chorionic vesicle, is situated within the ovary adjacent to the implantation cavity. It consists of a plicated lamina of large, lightly staining polygonal cells with faintly staining nuclei. On three aspects the lutein lamina is uniform in size. The lamina is infiltrated by connective tissue septa which are already vascularised. Within the lamina of lutein cells there is a delicate connective tissue layer enclosing a central blood clot. A second group of lutein cells form the walls of that part of the corpus luteum which is adjacent to the implantation cavity (fig. 7). These cells also constitute a plicated lamina which is oaly oneA Case of Early Human Ovarian Pregnancy 71
third of the thickness of the more extensive layer. Externally an uninterrupted layer of connective tissue separates the thinner lutein lamina from the implantation cavity. This layer contains decidua-like cells. Beyond each end of the lamina the internal and external layers of connective tissue become continuous with one another. Lutein cells are absent over considerable areas of these uniting bridges of connective tissue.
There is no direct relationship between the chorionic vesicle and the corpus luteum adjacent to it. The separation of the corpus luteum from the implantation cavity by a layer of connective tissue makes it unlikely that any lutein
of Yolk S
Amniotic y Vesicle
Barrier cell zone
Connective tissue lamina ,
Corpus Luteum entral Blood Mass in Corpus Fig. 7. Drawing from a photograph of a wax plate reconstruction of the chorionic vesicle and
embryonic papilla of H 153 showing the relations of surrounding parts. The corpus luteum is shown to be separated from the implantation cavity by a layer of connective tissue ( x 5).
cells are spread out in the wall of the gestation sac. Certainly they cannot be present in sufficient numbers to produce the zone of cells found bounding the intervillous spaces described as representing an ovarian decidua.
H 1538 is an addition to that class of ovarian pregnancy in which the chorionic vesicle lies wholly outside the corpus luteum. This condition is exhibited in the specimens described by Hewetson and Jordan-Lloyd (’06), Thompson (02), Webster (’04), Franz (’10), and Bryce, Teacher, and Kerr (’08). Whitehouse (’10) also described a specimen in which plasmodial masses were situated wholly outside a cystic corpus luteum. 72 John I. Hunter
The problem is to account for the extra-follicular position of the ovum in these cases.
A summary of the views of various authors in regard to the imbedding of the ovum is given by Graham (12). This writer is himself of the opinion that imbedding occurs in an epithelial lined tubule cut off from the fimbriated extremity of the Fallopian tube and that the ovum is attracted towards the corpus luteum. Webster (’04) believes that the fertilised ovum is imbedded in Mullerian duct inclusions which react to form a decidua, and that imbedding is extra-follicular. The wide distribution of ectopic decidua in cases of uterine pregnancy proves that such inclusions are unnecessary to provide a decidual reaction. Moreover, the large epithelial cells described by Graham and Webster are not consistently found in cases of ovarian pregnancy. In H 158 there is no justification for the view that the decidua cell barrier has arisen from Mullerian tissue. On the contrary, its connective tissue origin is displayed in those sections where the ovarian stroma directly bounds the intervillous spaces. In Hewetson and Lloyd’s specimen a layer of connective tissue completely separates the implantation cavity from the corpus Juteum as in H 153. The authors believe that the ovum was fertilised after extrusion from the Graafian follicle, and that the fertilised ovum then invaded the ovary at a higher level. They suggest that the grafting on the ovary may be due to the retention of the fertilised ovum in a crypt or by adhesions. Such cases are distinguished by Spiegelberg as epiovarian. Although the possibility of such an occurrence cannot be excluded it is unlikely.. However, the presence of inflammation of the ovary has been mentioned by Caturani as a possible aetiological factor in the production of ovarian pregnancy. Inflammatory processes of the ovary and in some instances of the tube in cases of ovarian _ pregnancy have been described by Caturani (14), Norris (709), Lockyer (717),
Young and Rhea (11) and Bandel (’02). No adhesions were to be found in the ease of H 153. Yet, in order to invade the ovary from without, it must be supposed that in some way the fertilised ovum was retained in the vicinity of the ovary until its trophoblastic layer was developed. Bryce, Teacher, and Kerr (08) while admitting the possibility of such an invasion in Hewetson and Jordan-Lloyd’s specimen suggest an alternative explanation for the occurrence of such cases in which connective tissue surrounds the implantation cavity. They believe that the ovum was primarily imbedded in the thin layer of connective tissue within the corpus luteum and that it burrowed through the wall of the corpus to reach its extra-follicular position. Although the chorionic vesicle was extra-follicular in the specimen described by Bryce, Teacher, and Kerr (’08), a gap was present in that part of the wall of the corpus luteum which adjoined the implantation cavity. A similar gap occurred in the specimen described by Franz (02). Bryce, Teacher, and Kerr state, in describing their specimen, that at the site of interruption of the corpus luteum, “the gestation sac is directly continuous with the interior of the corpus luteum by a band of tissue of doubtful character and more or less in a state of necrosis.” A Case of Early Human Ovarian Pregnancy 73
They interpret this strand as being comparable to the cone of fibrinous material occupying the point of entrance of the Bryce-Teacher ovum. Their conclusion is, therefore, that the ovum had been fertilised within the follicle and while only about -2 mm. in diameter migrated through the wall of the corpus luteum to become imbedded in the surrounding vascular connective tissue.
However, Bryce, Teacher, and Kerr admit a second possibility, viz. “that the ovum was arrested between the lips of the wound of the follicle, was there fertilised, and was then imbedded in the vascular stroma outside the follicle.” As they point out, this may explain Thompson’s case (’02) which lay “in the splayed out mouth of the follicle with no lutein tissue either on its free surfacé or between it and the corpus luteum.” However, since in their younger specimen the chorionic vesicle was wholly outside a corpus luteum they favour the first alternative. They state that if the vesicle had continued to grow the corpus luteum would have been invaded and destroyed as in Thompson’s case. They regard the migration theory as a more complete analogy to the conditions found in uterine implantation. They believe that all cases may be explained by assuming that they were primarily intra-follicular and only vary from one another in the degree of penetration of the connective tissue. They agree with van Tussenbroek that in her case the ovum imbedded itself excentrically in the connective tissue within the lutein lamina. If such an ovum progressed, the final result would again be destruction'of the corpus luteum. This explanation adheres to van Tussenbroek’s original definition that “ovarian pregnancy means pregnancy in an ovarian follicle.”
A study of H 153 which is at least in the same age period as the specimen described by Bryce, Teacher, and Kerr has convinced me that it is possible for the fertilised ovum to become primarily implanted in the connective tissue surrounding the mouth of the ruptured follicle. This is the alternative explanation which is dismissed by Bryce, Teacher, and Kerr. There is no cone of necrosed tissue joining the gestation sac with the interior of the corpus luteum in H 153 as in their specimen. Moreover, although the size of the lutein layer adjacent to the implantation cavity is reduced, it is nevertheless a well-formed plicated lamina. The reduction in size is readily explainable by the probability that the vascularisation of the area in which it is growing is interfered with due to the proximity of the implantation cavity.
The appearance in H 153 suggests that the ovum, together probably with cells of the discus proligerus still attached to it, was entrapped in the blood clot closing the wound of the ruptured follicle. Spermatozoa probably fertilised the ovum while lying in this situation as the coagulum formed. After the development of the trophoblast the ovum seems to have imbedded itself in the connective tissue surrounding the original opening and here development proceeded. In this case the growing ovum has formed for itself a capsule consisting of a thin outer layer of stroma, and a deep layer of connective tissue, separating the implantation cavity from the corpus luteum.
With the data presented by H 1538 the early ovarian pregnancy of Bryce,
Anatomy LvI 6 74 John I. Hunter
Teacher, and Kerr no longer occupies a unique position. In H 153 the cone of necrosed tissue described by Bryce, Teacher, and Kerr is not present. If their explanation that it is formed by the outward migration of the ovum were correct, it should be still visible in this specimen of corresponding age if this migration had taken place. Moreover, another explanation may be given to the necrosis of the lutein wall described by these authors, viz. that invasion of the corpus luteum by a growing ovum primarily implanted outside the follicle is in progress. They themselves state that “only at one point does the chorion come close to the ovarian stroma and that is opposite a gap in the wall of the corpus luteum.”
The further progress of development of a growing chorionic vesicle primarily imbedded outside the follicle would be identical with that described by Bryce, Teacher and Kerr for specimens which have secondarily invaded that tissue, viz. the corpus luteum will be invaded and finally will disappear. It is thus not unlikely, as suggested by Webster in 1904, that many cases obtained at a comparatively late stage of development described as follicular because of the presence of lutein cells in the gestation sac, may have been imbedded in the first instance in the thecal connective tissue and not in the follicle at all. In this connexion it is interesting to note that the extrafollicular specimens so far described are early specimens which suggest that extra-follicular imbedding is not uncommon. Probably, as suggested by Thomson (719), ovarian pregnancy may be produced by fertilisation of an ovum which has not been expelled from the follicle, and some cases are to be explained in this way. But it is certain that van Tussenbroek’s original definition which includes only such cases must be replaced by a more comprehensive statement to include primary imbedding outside the follicle as well.
SUMMARY AND CONCLUSIONS
The specimen described in this paper (H 153, Anat. Dept. Collection, Sydney University) has been shown from clinical and histological data to be an example of primary ovarian pregnancy of less than 20 days’ duration. The embryonic papilla (3-5 mm. in length) was enlarged by the infiltration of maternal blood into the yolk sac through the body stalk. The compressed chorionic vesicle measured 10 x 4-7mm. The specimen is of special value due to the early period of development. The following conclusions were derived from a consideration of its features.
(1) That a decidual reaction probably always occurs in early stages of ovarian impregnation.
(2) That the decidual tissue in H 153 is ploughed up by infiltration of the ovarian stroma by blood from the intervillous space; this infiltration has detached the villi from their attachments and probably led to retardation of growth of the chorionic vesicle and embryonic rudiment.
(8) That the ovum was fertilised as the closing coagulum was formed in the mouth of the Graafian follicle and proceeded to develop in the conA Case of Early Human Ovarian Pregnancy 75
nective tissue in this situation, imbedding itself ab initio outside the growing corpus luteum of pregnancy. It appears that ovarian pregnancy may result from the fertilisation and imbedding of an ovum within its follicle or at any point of its course in its progress outwards.
In conclusion I wish to express thanks to Messrs Schaeffer, Burfield and Bagnall of the University of Sydney for the technical assistance rendered me in the preparation of this account.
(01) Annine, G. P. and LirrLEwoop, H. “A Case of Primary Ovarian Pregnancy.” Trans. Obstets. Soc. xt, p. 14. London, 1901.
(10) Barrows, C.C. “Ruptured Ovarian Pregnancy.” Amer. Journ. Obstet. Lx, pp. 1078-1081. New York, 1910.
(12) Banks, A. G, “A Case of Ovarian Pregnancy.” Journ. Obstet. Gyn. Brit. Emp. xx, pp. 216-219. 1912.
(02) BanpeEL, R. Beitr. z. Klin. Chir. xxxvi, p. 657. 1902.
(796) Buanp Sutton. Surgical Diseases of the Ovaries and Fallopian Tubes, p. 35. 1896.
(701) Brit. Med. Journ. Vide Discussion of Anning and Littlewood’s specimen. 1901.
(04) BoxrsEBEEK. Monat. f. Geb. u. Gyn. Berlin, 1904. Vide Norris (’09).
(10) Bonp1, Joser. Wiener Klinische Wochenschrift, p. 774. 1910.
(08) Bryce, TEAcHER and Monro Kerr. Dev. and Imbedding of Human Ovum. An Early Ovarian Pregnancy. Glasgow, 1908.
(14) Carurant, M. “Ovarian Pregnancy with Report of a Case.” Amer. Journ. of Obstets. LXIx, pp. 409-428. 1914.
(13) Curene, G. “Ruptured very early Ovarian Pregnancy.” Obstet. Trans. Edin. xxxvm, p. 96. 1913. Also Edin. Med. Journ. x, p. 316.
(00) Crorr. “On an Anomalous Case of Extra-uterine Pregnancy probably Ovarian.” Trans. London Obstet. Soc. xu, p. 316.
(02) Franz, K. “Ueber Einbettung und Wachstum des Eies im Eierstock.” Hager’s Bettrige z. Geb. u. Gyn. vi. 1902. Vide Bryon, TEACHER and Kerr (’08).
(06) Freunp, H. W. and Taomt, R. Virchow’s Archiv f. Path. Anat. Bd cixxxm, Heft 1. 1906. Vide Ketiy and McIzroy (’06).
(14) Gms, A. E. and Looxyer, C. “A Case of Ovarian Pregnancy.” Proc. Roy. Soc. Med. vin, pp. 2-10. 1914-15.
(12) Granam, G.S. “A Case of Ovarian Pregnancy.” Journ. Med. Res. Xxv1, pp. 499-512. 1912.
(13) Grimspaxe, T. B. “Case of Ovarian Pregnancy with Full Time Foetus.” Journ. Obstet. and Gyn. of Brit. Emp. xxm, pp. 115-117. 1913.
(12) Hannes, W. “Ovarian-Graviditat.” Zeitschrift fiir Geburtsch. und Gyndkologie, LXxtI, pp. 269-271. 1912.
(09) Hastinas, Tweepy. “A Specimen of Ovarian Pregnancy.” Proc. Roy. Soc. Med. m, pt. 2. 1909-10. .
(06) Hewetson and Jorpan-Luoyp. Brit. Med. Journ. u, p. 568. 1906.
(11) Hotzanp, Earpury. “A Case of Ovarian Pregnancy probably Bilateral.” Journ. Obstet. and Gyn. of Brit. Emp. xx, pp. 295-298. 1911.
(21) Hunter, J. I. “A Case of Early Primary Ovarian Pregnancy.” Med. Journ. Australia, ut, No. 13, p. 233. 1921.
(08) Jacosson, 8. D. True Primary Ovarian Pregnancy. Contributions to the Science of Medicine and Surgery, New York Post Grad. Med. School and Hosp. p. 241. 1908.
(06) Ketty, J. K. and McItroy. Journ. Obstet. and Gyn. of Brit. Emp. 1x, p. 389. 1906.
(08) Kerr, J. M. M. “Case of Ovarian Pregnancy associated with an Intra-uterine Pregnancy.”’ Proc. Roy. Soc. Med. 1, No. 9, p. 268. 1908.
(10) Lea, A. W. W. “A case of Ovarian Pregnancy with Diffuse Intraperitoneal Haemorrhage.” Journ Obstet. and Gyn. of Brit. Emp. xv, pp. 182-187, 1910.
6—2 76 John I. Hunter
(12) Lenpon, A. A. Aust. Med. Gaz. xxx, p. 492. 1912.
(02) Leon, M. pg and Hotteman. Rev. de Gyn. June, 1902. Vide Ketity and McIzroy (’06).
(17) Lockyer, C. “Two Cases of Primary Ovarian Pregnancy.” Proc Roy. Soc. Med. x, No. 8, p. 169. June, 1917.
(15) Matt, F. P. “Tubal Pregnancy.” Surg. Gyn. Obstet. xx1, pp. 289-298. 1915.
- (’13) Matt, F. P. and Cutien, E. K. “An Ovarian Pregnancy located in the Graafian Follicle.”
Surg. Gyn. Obstet. xvm, pp. 698-703. 1913.
(02) Mayo, Rosson. “Primary Ovarian Gestation.” Trans. Lond. Obstet. Soc. xt1v, p. 215. 1902. Also Brit. Med. Journ. 11, p. 1447. 1902.
(08) Norris, C. C. and Mrrcuett, C. B. “Primary Ovarian Pregnancy with Report of a Case.” Surg. Gyn. Obstet. p. 460. May, 1908.
(09) Norris, C. C. “Primary Ovarian Pregnancy. The Report of a Case combined with Intrauterine Pregnancy.” Surg. Gyn. Obstet. 1x, p. 123. 1909.
(713) OLpFIELD, C. “Ovarian Gestation.” Journ. Obstet. Gyn. Brit. Emp. xxi, pp. 41, 42. 1913.
(21) Ray, H. M. “Primary Ovarian’ and Primary Abdominal Pregnancy: Their Morphological Possibility.” Surg. Gyn. Obstet. xxxu, No. 5, p. 437. 1921.
(78) SpPrecELBERG, Otto. Arch. f. Gyn. vin, p. 73. 1878. Also “Text-book of Midwifery,” New Sydenham Soc. Trans. p. 429.
‘(06) Taussic. “Ectopic Decidua Formation.” Surg. Gyn Obstet. p. 292. March, 1906.
(98) Taytor, J. W. “Ingleby Lectures.” Lancet, p. 1447. May, 1898.
(02) Tuompson, J. F. Trans. Amer. Gyn. Soc. xxvu, p. 338. 1902. Quoted by Ketiy and McIzroy (’06).
(719) Tomson, Artuur. “Rupture of the Graafian Follicle.” Journ. of Anat. tiv, Pt 1. 1919.
(99) van TUSSENBROEK, CATHERINE. Annales de Gyn. et Obstet. Dec. 1899. Quoted by Bryce, TEACHER, and KERR (’08).
(712) Verco, W. A. “Ovarian Pregnancy.” Aust. Med. Gaz. xxx1, p. 491. 1912.
(02) Wesster, J.C. Amer. Journ. Obstet. xiv, p. 302. 1902.
(704) —— “Study of Ovarian Pregnancy.” Amer. Journ. Obstet. L, pp. 28-50. 1904.
(707) Trans. Amer. Gyn. Soc. xxx, p. 122. 1907.
(10) Wastrexovss, B. “Remarks on a case of Probable Ovarian Pregnancy.” Brit. Med. Journ. I, pp. 1276-1280. London, 1910.
(08) Wiiu1ams, Wuitriper. Kelly and Noble’s Gynaecology, 1908.
(17) —— Tezt-book of Obstetrics, 1917.
(11) Youne, E. B. and Ruga, L. J. “Ovarian Pregnancy: Report of a Case.” Boston Medical and Surgical Journ. OLXIv, p. 264. 1911.
[* Note. Since this article has been in the press Contributions to Embryology, Vol. x11, No. 56, Washington, 1921, which includes “Ovarian Pregnancy”? by F. P. Mall and A. W. Meyer, has become available.]