Placenta
Volume 31, Issue 4 , Pages 269-276, April 2010

Placental markers of twin-to-twin transfusion syndrome in diamniotic–monochorionic twins: A morphometric analysis of deep artery-to-vein anastomoses

  • M.E. De Paepe

      Affiliations

    • Department of Pathology, Women and Infants Hospital, 101 Dudley Street, Providence, RI 02905, USA
    • Department of Pathology and Laboratory Medicine, Alpert Medical School of Brown University, Providence, RI, USA
    • Corresponding Author InformationCorresponding author. Department of Pathology, Women and Infants Hospital, 101 Dudley Street, Providence, RI 02905, USA. Tel.: +1 401 274 1122x1544; fax: +1 401 453 7681.
    • ,
  • S. Shapiro

      Affiliations

    • Department of Pathology, Women and Infants Hospital, 101 Dudley Street, Providence, RI 02905, USA
    • ,
  • D. Greco

      Affiliations

    • Department of Pathology, Women and Infants Hospital, 101 Dudley Street, Providence, RI 02905, USA
    • ,
  • V.L. Luks

      Affiliations

    • Department of Pathology, Women and Infants Hospital, 101 Dudley Street, Providence, RI 02905, USA
    • ,
  • R.G. Abellar

      Affiliations

    • Department of Pathology, Women and Infants Hospital, 101 Dudley Street, Providence, RI 02905, USA
    • Department of Pathology and Laboratory Medicine, Alpert Medical School of Brown University, Providence, RI, USA
    • ,
  • C.H. Luks

      Affiliations

    • Department of Pathology, Women and Infants Hospital, 101 Dudley Street, Providence, RI 02905, USA
    • ,
  • F.I. Luks

      Affiliations

    • Program in Fetal Medicine, Alpert Medical School of Brown University, Providence, RI, USA

Accepted 22 December 2009. published online 11 January 2010.

Article Outline

Abstract 

Twin-to-twin transfusion syndrome (TTTS) is a multifactorial disorder that develops in 9–15% of diamniotic–monochorionic twin gestations. While the pathogenesis of TTTS remains poorly understood, unbalanced deep artery-to-vein (AV) anastomoses have traditionally been implicated in the gradual shift of blood from donor to recipient. The aim of this study was to define the placental markers of twin-to-twin transfusion syndrome, with special emphasis on the deep AV anastomoses. A prospective cohort of 284 consecutive diamniotic/monochorionic twin placentas was examined at Women and Infants Hospital between 2001 and 2008. Following exclusion of monoamniotic, multiple, disrupted and laser-treated placentas, 218 twin placentas (21 TTTS and 197 non-TTTS controls) formed the subject of this study. Placentas were injected with color-coded dyes. Anatomic characteristics and choriovascular anastomotic patterns of TTTS placentas were compared with non-TTTS controls. The TTTS placentas showed significantly higher frequencies of velamentous cord insertion, magistral vascular distribution patterns, uneven placental sharing, absence of AA anastomoses and presence of VV anastomoses. Deep AV anastomoses were identified in ≥95% of TTTS and non-TTTS placentas and were overall more abundant than previously reported. The total and net numbers of AV anastomoses were similar in both groups. However, the net cross-sectional area of AV anastomoses, which also takes into account the caliber of the vessels, was significantly smaller in TTTS placentas. There was no correlation between the direction of the AV imbalance and the twin donor/recipient status. In conclusion, TTTS has distinct placental characteristics, warranting their routine inclusion in the diamniotic–monochorionic placental pathology report. Our findings suggest imbalance of AV anastomoses is not required for the development for TTTS, although their presence, whether balanced or unbalanced, may contribute to the creation or perpetuation of the syndrome. Elucidation of the role of the various placental determinants in diamniotic–monochorionic twin gestations may lead to further refinement of therapeutic strategies.

Keywords: Monochorionic twin, Placenta, Vasculature, Anastomoses, Twin-to-twin transfusion syndrome

 

Back to Article Outline

1. Introduction 

Approximately 9–15% of monochorionic twin gestations are complicated by severe chronic twin-to-twin transfusion syndrome (TTTS), characterized by a gradual shift of blood from the donor twin to the recipient through placental vascular connections between the fetuses [1], [2], [3]. The prognosis of untreated chronic TTTS, diagnosed in mid-trimester, is extremely poor and associated with morbidity and mortality rates exceeding 70% [4]. In view of the poor outcome of TTTS, aggressive treatment modalities have been developed, including umbilical cord ligation, repetitive amnioreduction of the polyhydramniotic sac and fetoscopic laser coagulation of the communicating vessels. The latter technique, first described by De Lia et al. in 1990 [5], uses a laser beam to photocoagulate the intertwin vascular communications believed to be responsible for TTTS.

The pathogenesis of TTTS remains incompletely understood. While the vast majority of diamniotic–monochorionic twin gestations have intertwin anastomoses, only a fraction of them develops TTTS. The risk for development of TTTS in diamniotic–monochorionic twin pregnancies has been linked to a number of anatomic placental characteristics, including uneven sharing of the single-disc placenta and velamentous or marginal cord insertion [6], [7], [8], [9], [10], [11], [12].

In view of its alleged critical role in the pathogenesis of TTTS, the architecture of the intertwin vascular connections has been the subject of numerous studies [11], [13], [14], [15], [16]. Placental vascular communications occur in virtually all monochorionic placentas [11], [16], [17] and are either superficial (lying on the surface of the chorionic plate) or deep (within the placental parenchyma). Artery-to-artery (AA) and vein-to-vein (VV) anastomoses are direct and superficial twin–twin connections. In contrast, artery-to-vein (AV) anastomoses are deep and refer to a shared cotyledon with arterial supply from one twin and venous drainage to the other. Unlike the superficial AA and VV anastomoses, AV anastomoses are obligatory unidirectional and AV imbalance directed from donor to recipient is traditionally believed to be a critical factor in the uncompensated flow from one twin to the other [18].

Previous studies of the placental angioarchitecture in TTTS pregnancies have mainly focused on the superficial AA and VV anastomoses [12], [19], [20]. In spite of their alleged critical role in the pathogenesis of TTTS, AV anastomoses have received little attention in the literature. This relative neglect of the AV anastomoses may be attributable to the fact that their study is more complex and time-consuming than that of other placental anatomic variables. Indeed, the type of cord insertion, degree of placental sharing and even presence or absence of superficial AA and VV anastomoses can usually be visualized by standard gross examination of the placenta. In contrast, accurate quantitative analysis of the AV anastomoses requires relatively laborious and specialized vascular injection techniques [16].

The aim of this study was to perform a detailed prospective morphometric analysis of a large consecutive series of monochorionic twin placentas in order to determine the frequency of candidate placental markers in TTTS placentas and, in particular, to determine the potential role of unbalanced deep AV anastomoses. To facilitate identification and morphometric analysis of the AV anastomoses, we used our previously described vascular injection technique [16], which allows accurate assessment of number, direction and caliber of the vessels and thus provides an anatomic proxy for net AV flow (im)balance. Increased knowledge of the placental anatomy and angioarchitecture in diamniotic–monochorionic twin gestations may lead to a better understanding of the pathophysiology of TTTS, aid in the further refinement of therapeutic strategies, and eventually be of prognostic value in this often lethal condition.

Back to Article Outline

2. Materials and methods 

A consecutive series of 284 monochorionic twin placentas was submitted to the Department of Pathology at Women and Infants Hospital between 2001 and 2008. This study focused on the placental markers and angioarchitecture of TTTS in diamniotic–monochorionic twin placentas. Triplet and quadruplet placentas, monoamniotic placentas, placentas of pregnancies complicated by twin reversed arterial perfusion (TRAP) sequence and placentas with remote (>48h prior to delivery) demise of one twin were excluded. In addition, placentas in which cord avulsion of partial disruption of the chorionic plate vasculature precluded reliable vascular injection studies were excluded. The accompanying charts were reviewed to determine whether the pregnancy was complicated by chronic TTTS, as defined by ultrasonographic evidence of severe polyhydramnios in one twin and concomitant oligohydramnios in the other, in some cases associated with collapse of the bladder in the donor twin, critically abnormal Doppler studies, hydrops and/or demise of one or both twins [21]. Laser-treated TTTS placentas were excluded from the study.

The placental weight was compared with reported institutional reference values for age-matched twin placentas [22] and classified as small (<10th centile), large (>90th centile) or appropriate for gestational age. Gross examination of the placenta and injection of the placental vasculature were performed as previously described in detail [16], [23].

The type of cord insertion (central/paracentral, marginal, or velamentous), the number of umbilical arteries, and the relative distribution of the vascular territories were noted. Velamentous cord insertion was defined as cord insertion into the fetal membranes rather than onto the placental disc. Marginal cord insertion was defined as cord insertion at the edge of the placental disc. In addition to these routine twin placental parameters, we recorded the vascular distribution patterns of each twin. The vascular distribution patterns were categorized as disperse, magistral or mixed, as previously described [23]. The disperse pattern was defined as a superficial vascular arrangement characterized by regular, near-symmetric dichotomous branching, resulting in a progressive diminution of vascular caliber. For the purpose of this study, a pattern was categorized as disperse if ≥75% of the twin's vascular territory showed this pattern. A vascular pattern was categorized as magistral if ≥75% of the vascular territory showed relatively large-sized vessels, extending from the insertion of the cord to the periphery without a significant reduction of diameter. The mixed vascular distribution pattern was defined as the presence of combined disperse and magistral-type vessels, with each type involving less than 75% of the individual twin's territory.

The presence, type, number and caliber of intertwin anastomoses (AA, VV or AV) were recorded by a single observer (MEDP) based on placental examination at the time of vascular injection. The vascular diameters were determined with a caliper and/or ruler. Superficial AA and VV anastomoses were identified as direct superficial communications between two homonymous umbilical vessels. Arteries were recognized by their general tendency to cross over corresponding veins. Deep AV anastomoses were identified where an unpaired artery from one twin was seen penetrating the chorionic plate close to (<1.0cm) an unpaired vein from the other twin. Deep AV anastomoses were categorized by direction of flow. For AA and VV types, the caliber of the anastomosis was measured as the minimum external diameter along the course of the anastomosis. The diameter of an AV anastomosis was defined as the diameter of the feeding artery at its narrowest point. The cross-sectional area (CSA) of each AV anastomosis was calculated (CSA=πr2 where r=radius of feeding artery) and given a positive or negative value, depending on the direction of the flow (positive if donor to recipient, negative if recipient-to-donor). Thus, the net cross-sectional surface area (NCSA) represents the sum of the cross-sectional surface areas of all AV anastomoses in one direction minus the sum of all surface areas of AV anastomoses in the other direction, expressed as an absolute value.

Finally, we determined the relative placental share of each twin, based on the distribution of respective chorionic vessels. All injected placentas were photographed. Each vascular territory was mapped by digital image analysis by manually tracing the margins demarcated by the presence of dye and expressed as a percentage of the overall surface area. Analysis of the placentas was performed by a single perinatal pathologist (MEDP) who had no knowledge of the clinical course.

Following study of the gross placental and vascular anatomy, routine histologic sections were prepared, including umbilical cords, sections of marginal and dividing membranes, representative villous parenchyma of both twins, and any focal lesions. Description of the microscopic findings is beyond the scope of the present study. Of note, chorionicity was confirmed by microscopic examination of the dividing membrane in all cases.

Data were compared by chi-square, Fisher exact, Mann–Whitney and Student t-tests, where applicable. A P value <0.05 was considered statistically significant. The study was approved by the Institutional Review Board.

Back to Article Outline

3. Results 

3.1. Patient population and clinical data 

A cohort of 284 consecutive monochorionic placentas was examined between 2001 and 2008. Triplet and quadruplet placentas (11), monochorionic–monoamniotic placentas (11), placentas of gestations complicated by twin reversed arterial perfusion (TRAP) sequence (2) and placentas with remote (>48h prior to delivery) fetal demise of one twin (7) were excluded. Of the remaining 253 diamniotic–monochorionic twin placentas, 53 were complicated by TTTS, based on clinical and ultrasound evidence. Of the 53 TTTS cases, 29 were treated by laser coagulation of communicating vessels. These laser-coagulated TTTS placentas were excluded from this study, as the number of grossly identifiable anastomoses has been shown to be significantly reduced after laser-ablation [24].

Twenty-four non-laser-treated TTTS cases, denoted as “TTTS” are the subject of this study. The findings in the 24 TTTS placentas were compared with 200 placentas of non-TTTS diamniotic–monochorionic twin gestations, described as “non-TTTS” control placentas. The gestational age at delivery was lower for TTTS twins than for non-TTTS control twins (mean gestational age 27.3 weeks versus 34.5 weeks) (Table 1). Accordingly, the birth weights of the TTTS twins were lower than those of non-TTTS twins (Table 1). Half (12/24) of the TTTS twins showed ≥20% birth weight discordance, compared with 15% of non-TTTS twins. Thirty-three percent of TTTS gestations were complicated by intrauterine fetal demise, most often of both twins. In contrast, only 2% of non-TTTS pregnancies were associated with fetal demise.

Table 1. Patient data.
TTTS twins (24)Non-TTTS twins (200)P
Gest. age (weeks)27.3±3.8 (20–33)34.5±3.8 (21–42)<0.01
BW larger twin (g)1149±525 (250–1935)2271±697 (579–3845)<0.01
BW smaller twin (g)904±449 (190–1527)2008±676 (520–3523)<0.01
Difference BW (%)24.0±14.0 (3–67)12.0±8.8 (0–18)<0.01
Fraction with BW discord. ≥20%12/24 (50%)30/200 (15%)<0.001
Associated with IUFD8/24 (33%)4/200 (2%)<0.001
IUFD of both twins6/24 (25%)2/200 (1%)<0.001
IUFD of one twin2/24 (8%)2/200 (1%)0.08

Values represent mean±SD (range) or proportion (%) of (N) twins.

BW: birth weight; IUFD: intrauterine fetal demise.

This study focused on the vascular anastomotic patterns of non-laser-treated, ‘virgin’, placentas and was performed in an institution that only offers laser treatment for advanced TTTS (persistent Quintero stage II and above). Inherently, therefore, there was a difference between the laser-treated cases excluded from this study and non-laser-treated TTTS cases included in this study. All laser-treated TTTS cases had reached stages II and above at the time of surgery. The non-lasered TTTS cases included a similar group of advanced TTTS, who did not undergo surgery for a range of clinical scenarios. The reasons for not undergoing laser surgery for TTTS varied from unrecognized TTTS (particularly for placentas sent to us from outside institutions), advanced gestational age (upper gestational limit for laser surgery at our institution is 26 weeks), refusal of patients to undergo surgery, fetal demise before laser surgery could be performed, and technical impossibility to perform laser surgery (due to lack of a placenta-free window by ultrasound). In addition, about one-third of the non-laser-treated TTTS group consisted of stage I TTTS cases.

3.2. General placental anatomy 

The placental weights, determined after removal of membranes and cords, were compared with institutional reference values for twin placentas [22]. In accordance with the different age distribution of the two study groups, the placental weights of TTTS twins were significantly lower than those of non-TTTS control twins (Table 2). Almost 90% of non-TTTS placentas were of appropriate weight for gestational age compared with 71% in the TTTS group (Table 2).

Table 2. Placental and choriovascular anatomy.
GENERAL ANATOMY
TTTS placentas (24)Non-TTTS placentas (200)P
Placental weight (g)471±161708±210<0.01
Placenta AGA17/24(71%)172/200(86%)0.06
Placenta LGA4/24(17%)11/200(6%)0.06
Placenta SGA3/24(13%)14/200(7%)NS
Velamentous cord insertiona14/42(33%)41/394(10%)<0.0001
Marginal cord insertiona8/42(19%)82/394(21%)NS
Velamentous or marginal cord insertiona22/42(52%)123/394(31%)<0.01
Magistral/mixed vascular distribution18/21(86%)118/197(60%)<0.02
>25% difference placental share11/21(52%)50/197(25%)<0.01
Single umbilical artery3/48(6.3%)10/394(2.5%)NS
CHORIOVASCULAR ANATOMY
TTTS placentas (21)Non-TTTS placentas (197)P
Total # anastomoses6(2–23)7(0–23)NS
AA anastomoses present12/21(57%)175/197(89%)<0.001
VV anastomoses present8/21(38%)28/197(15%)<0.02
Total # deep AV anastomoses5(0–22)6(0–22)NS
Net # deep AV anastomoses2(0–7)2(0–15)NS
NCSA (mm2)0.39(0–6.67)0.98 (0–42.78)<0.05
Fraction with NCSA>1mm2 (%)7/21(33%)91/197(46%)NS
Fraction with NCSA>5mm2 (%)1/21(5%)26/197(13%)NS

Values represent mean±SD (range) or proportion (%) of (N) twins.

AGA: appropriate for gestational age; LGA: large for gestational age; SGA: small for gestational age. NCSA: net cross-sectional area of AV anastomoses.

aType of cord insertion per individual fetus.

The TTTS placentas were more likely to have velamentous insertion compared with non-TTTS control placentas (33% versus 10%, P<0.0001). The frequency of marginal cord insertion was similar in both groups (19% versus 21%). In agreement with our previous observations [23], magistral or mixed magistral/disperse vascular distribution patterns were more prevalent in TTTS placentas than in non-TTTS controls (86% versus 60%, P<0.02). Further, markedly uneven placental sharing, defined as >25% difference in placental territory between twins, occurred more than twice as frequently in TTTS placentas than in non-TTTS controls (52% versus 25%, P<0.01). The smaller side corresponded to the donor side in all TTTS cases with uneven placental share. A single umbilical artery tended to be seen more frequently in TTTS placentas, although this difference did not reach statistical significance (6.3% versus 2.5%). When present, the single umbilical artery was always noted in the cord of the donor twin.

3.3. Choriovascular anatomy: superficial anastomoses (AA and VV) 

The intertwin vascular anastomoses were categorized (AA, VV or AV) and quantified following color-coded dye injection of the placental vessels, as previously described [16]. The vascular anastomoses could be assessed in 21 TTTS and 197 non-TTTS cases. In the remaining cases, partial or complete avulsion of an umbilical cord, resulting in focal disruption of the chorionic plate vasculature, prevented reliable vascular injection studies and interpretation of the anastomotic patterns. The total number of anastomoses visualized in the chorionic plate varied widely between cases and reached a maximum of 23 in both TTTS and non-TTTS control groups. All TTTS placentas had at least one intertwin anastomosis, whereas 5/197 (2.5%) non-TTTS placentas displayed a complete absence of vascular communications. Superficial AA anastomoses were significantly more frequent in non-TTTS placentas than in TTTS (89% versus 57%, P<0.001). None of the TTTS placentas had more than one AA anastomosis. Among the non-TTTS cases, two placentas had 2 and one had 3 AA anastomoses. The diameter of AA anastomoses ranged from 0.05 to 0.4cm in TTTS placentas (mean: 0.175±0.097cm) and from 0.05 to 0.6cm in non-TTTS controls (mean: 0.174±0.104cm).

Vein-to-vein anastomoses were the least prevalent type of intertwin anastomoses. These anastomoses were more than twice as frequent in TTTS placentas as in non-TTTS controls (38% versus 15%, P<0.002). None of the TTTS placentas had more than one VV anastomosis, whereas six non-TTTS control placentas had 2 VV anastomoses. The maximal diameter of VV anastomoses was 0.6cm in TTTS placentas (mean: 0.263±0.185cm) and 0.4cm in non-TTTS controls (mean: 0.207±0.110cm).

3.4. Choriovascular anatomy: deep AV anastomoses 

Deep AV anastomoses were characterized by the dipping of an unpaired artery from one twin into the placenta close to a corresponding unpaired vein from the opposite twin (Fig. 1). AV anastomoses were seen in 20/21 (95%) TTTS placentas and 189/197 (96%) of non-TTTS placentas. The total number of deep AV anastomoses (both directions combined) was comparable in both groups (median number of AV anastomoses: 5 in TTTS and 6 in non-TTTS control). The total number of deep AV anastomoses varied greatly and reached a maximum of 22 in both groups. The single TTTS placenta showing a complete absence of deep AV anastomoses had, in addition to a 2mm-wide AA anastomosis, a relatively large, 4mm-wide VV anastomosis, marginal insertion of one cord, and a magistral vascular pattern. The placental sharing was relatively even (10% difference).

  • View full-size image.
  • Fig. 1 

    Demonstration of angioarchitecture of diamniotic–monochorionic twin placentas following vascular injection. A. Representative diamniotic–monochorionic placenta following vascular injection. Color code: Artery left twin: yellow; vein left twin: green; artery right twin: yellow (through AA anastomosis); vein right twin: black. Arteries cross over veins. Arrows indicate AA anastomosis. B. Close-up of (A) showing abundant AV anastomoses from left twin (A, yellow) to right twin (V, black), indicated by white arrow heads. A minuscule right-to-left AV anastomosis is shown by black arrow head (bottom). A and B: Non-TTTS-pregnancy, 30 weeks' gestation, history of premature rupture of membranes. C. Placenta of pregnancy complicated by TTTS. Left twin is donor, right twin is recipient. Color code: Artery left twin: red; vein left twin: yellow; artery right twin: red (through AA anastomosis); vein right twin: green. Velamentous cord insertion and small placental share of left (donor) twin are noted, as well as relative paucity of intertwin anastomoses. Arrows indicate AA anastomosis. D. Close-up of (C) demonstrating several AV anastomoses from right (recipient) twin to left (donor) twin (black arrow heads). C and D: TTTS-pregnancy, 32 weeks' gestation, fetal demise of both twins.

In addition to the total number of AV anastomoses (both directions), we determined the net number of AV anastomoses, defined as the absolute value of the number of AV anastomoses in one direction subtracted from the number of anastomoses in the other direction. The median net number of AV anastomoses was 2 in both groups, with values ranging from 0 to 7 in TTTS and from 0 to 15 in non-TTTS placentas (Table 2). Interestingly, high net numbers of AV anastomoses were more typically seen in non-TTTS placentas than in TTTS placentas. For instance, none of the TTTS but 8/197 (4%) non-TTTS placentas had a net number of deep AV anastomoses higher than 7 (Table 2).

We further determined the net cross-sectional area (NCSA) of deep AV anastomoses. The NCSA takes into account not only the number and direction, but also the caliber of AV anastomoses, and thus may serve as closer anatomic proxy for net blood flow through deep AV anastomoses. Surprisingly, the NCSA was significantly smaller in TTTS placentas than in non-TTTS controls (median NCSA: 0.39mm2 in TTTS versus 0.98mm2 in non-TTTS placentas, P<0.05). In 14/21 (67%) of TTTS placentas the AV imbalance was minimal (between 0 and 1mm2). Strikingly, in 7/21 (33%) TTTS placentas, there was no evidence of any AV imbalance (NCSA=0). The NCSA ranged from 0 to 6.7mm2 in TTTS placentas and up to 42.8mm2 in non-TTTS placentas. Only one TTTS case but 26/197 (13%) non-TTTS cases had an NCSA larger than 5mm2 (Table 2).

As shown in Fig. 2, there was no significant correlation between net number of AV anastomoses and NCSA, indicating that assessment of vessel number alone may not offer an accurate estimate of AV imbalance. Fig. 2 further illustrates the higher proportion of cases with large numbers of net AV anastomoses and large NCSA in non-TTTS versus TTTS placentas.

3.5. Correlation between presence/absence of superficial vessels and frequency of TTTS in placentas with AV imbalance 

As noted above, AA anastomoses were present in 187/218 (86%) placentas. About half (93/187) of the placentas with AA anastomoses showed an AV imbalance. The prevalence of TTTS in placentas with AA anastomoses was equally low whether the AV anastomoses were balanced or not (9% versus 4%). In the absence of AA anastomoses, the rate of AV imbalance was lower than in the presence of AA anastomoses (5/31 (16%) versus 93/187 (50%), P<0.001). The prevalence of TTTS in placentas without AA anastomoses was similar in AV-balanced and AV-unbalanced placentas (23% versus 60%, difference not significant).

In placentas with AV imbalance, the frequency of TTTS was higher in the absence of AA anastomoses than in presence of AA anastomoses (60% versus 4%, P<0.001) (Fig. 3). Notably, a similar difference was noted in AV-balanced placentas, where the frequency of TTTS was also significantly higher in placentas lacking AA anastomoses (23% versus 9%, P<0.05).

  • View full-size image.
  • Fig. 3 

    Frequency of TTTS in AV-balanced and unbalanced placentas in association with superficial AA and VV anastomoses. *P<0.05; **P<0.02; ***P<0.001 versus same type of AV balance in presence of corresponding superficial anastomoses.

We then determined whether the presence or absence of VV anastomoses correlated with the occurrence of TTTS in placentas with AV imbalance. In AV-unbalanced placentas, the proportion of cases with TTTS was higher if a VV anastomosis was present than if it was absent (17% versus 5%), while an opposite tendency was noted in AV-balanced placentas (28% versus 50%). However, none of these differences were statistically significant.

3.6. Correlation between direction of AV imbalance and donor/recipient status in TTTS placentas 

Unbalanced flow through deep AV anastomoses in twins with TTTS is generally believed to be directed preferentially from donor to recipient. To test the validity of this concept, we determined the direction of the AV NCSA in TTTS placentas with unbalanced AV anastomoses with respect to donor/recipient status. In 4/7 TTTS placentas with NCSA>1mm2, the direction of deep AV imbalance, as estimated by the sense of the NCSA, was from donor to recipient. However, in the remaining three, the AV imbalance was directed from recipient-to-donor. These findings suggest that the direction of the AV flow imbalance is not a critical determinant of the donor/recipient relationship in TTTS twins. Of note, the donor side always correlated with the smaller placental share in unevenly shared placentas. Accordingly, the AV imbalance was directed from smaller to larger placental portion in four of seven AV-unbalanced cases, and from larger to smaller in three.

3.7. Sensitivity and specificity of key placental markers as post-hoc predictors of TTTS 

To evaluate the validity of placental examination as post-hoc predictor of TTTS, we determined the sensitivity, false positive rate and positive predictive value of the main candidate placental TTTS markers. As shown in Table 3, none of the individual candidate placental variables are reliable post-hoc predictors of TTTS. For instance, while demonstration of velamentous/marginal cord insertion or magistral/mixed vascular distribution patterns will allow detection of >85% of TTTS cases, these findings have low specificity, resulting in high false positive rates (53% and 60%, respectively). Conversely, absence of AA anastomoses is not a sensitive parameter for the presence of TTTS (sensitivity of 43%), but it has a higher specificity and, accordingly, a lower false positive rate (11%). As noted before, AV balance or imbalance alone did not correlate with TTTS (false positive rate approaching 50%).

Table 3. Validity of placental examination as post-hoc TTTS test.
Detection rate (Sensitivity)SpecificityFalse positive rate (1 – specificity)
Velamentous/marginal cord insertion86%47%53%
Magistral/mixed vascular distribution86%40%60%
>25% difference placental share52%75%25%
Single umbilical artery6%97%3%
AA absent43%89%11%
VV present38%86%14%
AV imbalance33%54%46%
Velamentous/marginal cord insertion+AA absent43%92%8%
Velamentous/marginal cord insertion+>25% difference placental share43%88%12%
Velamentous/marginal cord insertion+VV present29%92%8%
AA absent+>25% difference placental share10%96%4%
AA absent+AV imbalance14%99%1%

AA: artery-to-artery; VV: vein-to-vein; AV: artery-to-vein.

We then determined the value of the most promising combinations of placental variables as predictors of TTTS (Table 3). While combining the finding of velamentous/marginal cord insertion with other key variables resulted in decreased sensitivity, the specificity increased to acceptable levels (88%–92%), which resulted in relatively low false positive rates (8%–12%). The highest specificity (99%) was achieved by combining the absence of AA anastomoses with the presence of AV imbalance. However, this combination was not often present (sensitivity or detection rate of 14%). It needs to be emphasized that the validity of the post-hoc analysis for these various combinations is limited in view of the relatively small number of TTTS placentas.

Back to Article Outline

4. Discussion 

We determined placental markers of TTTS in a large cohort of monochorionic twin placentas examined over an 8-year period in a single institution. In addition to recording placental anatomic characteristics that can be evaluated by routine gross examination, we performed a detailed morphometric analysis of the choriovascular architecture with special attention to the deep AV anastomoses. In accordance with previous studies [7], [8], [9], [11], [12], [14], [15], [20], we found that placentas from pregnancies complicated by TTTS (henceforth termed “TTTS placentas”) were significantly more likely to have uneven placental sharing, absence of AA anastomoses and presence of VV anastomoses. This larger study further confirms the higher prevalence of magistral or mixed magistral/disperse vascular distribution patterns in TTTS placentas reported previously [23].

We also determined the incidence of velamentous and marginal cord insertion in TTTS and non-TTTS placentas. While both peripheral insertion types are considered aberrations of the normal (para) central type of insertion, they exhibited a distinct relationship with TTTS. In this study, the frequency of velamentous cord insertion was significantly higher in TTTS placentas than in non-TTTS placentas (33% versus 10%), whereas the frequency of marginal cord insertion was similar in both groups (19% versus 21%). The higher frequency of velamentous cord insertion in TTTS placentas is in agreement with several other studies [8], [10], [25], [26]. Interestingly, however, Bajoria et al. [14] and Lopriore et al. [27], reported similar frequencies of velamentous insertion in TTTS and non-TTTS placentas. The reason for these discrepant findings remains undetermined. A previous report from our institution [23] described similar frequencies of peripheral (velamentous or marginal) cord insertions in TTTS and non-TTTS placentas as well, which may be due to selection bias in this much smaller study (comprising only six non-lasered TTTS placentas).

The primary aim of this study was to perform a detailed morphometric analysis of the deep AV anastomoses, which are traditionally believed to be critical factors in the pathogenesis of TTTS. Two vascular injection-based approaches were used to estimate the degree of AV flow imbalance. First, we determined the net number of AV anastomoses, defined as the absolute value of the number of AV anastomoses in one direction subtracted from the number of AV anastomoses in the other direction. This measure does not take into account the size of the AV anastomoses, which was found to vary greatly between placentas and even within the same placenta. We therefore introduced assessment of the absolute net cross-sectional area (NCSA) as a more accurate proxy for AV flow balance between the twin circulations. Of note, we found a poor correlation between net number of anastomoses and NCSA, emphasizing the importance of considering vessel size in studies of placental angioarchitecture.

Several observations were made with regard to the prevalence and associations of deep AV anastomoses. First, the average number of AV anastomoses in this study was much higher than has been reported during fetoscopy in vivo [28] or in vascular injection studies ex vivo [12], with a maximum of 22 AV anastomoses per placenta detected in both TTTS and non-TTTS groups. We speculate that the vascular injection technique used [16], characterized by diligent ‘milking’ of the chorionic plate vessels in preparation of the injection and use of a dye mixture of relatively low viscosity, may have contributed to visualization and identification of even the smallest hair-sized vessels. It needs to be emphasized that laser-treated placentas were excluded from this study to avoid underestimation of the intertwin communications.

Second, we determined that AV anastomoses were present in virtually all diamniotic–monochorionic twin placentas (95% of TTTS and 96% of non-TTTS placentas). Third, we determined that the total number and net number of AV anastomoses were similar in both groups, in agreement with other studies [12], [29]. However, the NCSA, which also considers the caliber of the anastomoses and thus more accurately estimates AV flow, was significantly smaller in TTTS placentas compared with non-TTTS placentas. These findings thus suggest, counterintuitively, that the degree of AV imbalance in TTTS placentas is lower than that seen in non-TTTS placentas.

Fourth, we determined that AV imbalance was not essential for the development of TTTS. Two thirds of TTTS cases showed no or minimal AV imbalance, as judged by NCSA<1mm2. More strikingly, there was no morphometric evidence of any AV imbalance at all (NCSA=0mm2) in one-third of TTTS cases. Interestingly, we found that non-TTTS placentas were more likely to show marked degrees of AV imbalance than TTTS placentas. Very large numbers of AV anastomoses (total number>10, net number>8) were almost exclusively seen in non-TTTS placentas. Similarly, large NCSA values (>10mm2) were limited to non-TTTS placentas, reaching a striking maximum of 43mm2 in the most extreme case.

Finally, we determined that there was no correlation between the direction of the AV imbalance and the donor-versus-recipient status of TTTS placentas. Indeed, the direction of AV imbalance was as likely to be directed from donor to recipient as from recipient-to-donor. In this study, the donor side was the smaller placental part, and the recipient's the larger in all TTTS cases with uneven placental sharing.

The results from our morphometric analyses suggest that anatomic, postpartum AV imbalance per se is not essential for development of TTTS syndrome in diamniotic–monochorionic placentas. It is important to note that cross-sectional area of a blood vessel only correlates with blood flow if it can be assumed that pressure differential, blood vessel diameter and even viscosity are constant, which is likely not the case in vivo. While imbalance of AV anastomoses may not be required, there is some evidence that the presence of AV anastomoses may contribute to the development – and/or maintenance – of the syndrome. We speculate that the functional role of AV anastomoses in TTTS may be permissive rather than causative, and possibly related to their potential to allow transfer of soluble vasoactive substances between twins. Soluble factors that have been shown to cause, sustain or worsen TTTS include brain natriuretic peptide, endothelin-1 [30], and elements of the renin–angiotensin system [31], [32].

The view that AV anastomoses are essential for TTTS needs to be reconciled with the finding that TTTS can occur, albeit exceedingly rarely, in the absence of visible AV anastomoses. In our study, 1/21 TTTS placentas failed to show a deep AV anastomosis, demonstrable on the chorionic plate after vascular injection. Although unconfirmed in this study, we speculate that intraparenchymal deep AV anastomoses may have existed in our case, as recently described [33]. While considered harmless by some [34], it is conceivable that these intraparenchymal anastomoses, not visible by inspection of the chorionic plate, may occasionally contribute to the pathogenesis of TTTS.

In concordance with several other studies [7], [11], we found that superficial AA anastomoses were significantly more common in non-TTTS placentas than in TTTS placentas. The relative paucity of AA anastomoses in TTTS placentas has been universally interpreted as evidence of a protective role for AA anastomoses. According to this deeply vested notion, AA anastomoses are believed to play a functional role in TTTS by compensating for any hemodynamic imbalances created by uneven AV anastomoses [11], [19]. This model of AA anastomoses having a protective role has possibly been supported by mathematical models [19].

While the lower prevalence of AA anastomoses in TTTS placentas is indisputable, we would like to urge caution in interpreting this association as evidence of a functional or even causative relationship. It is possible that the frequent absence of AA anastomoses, as well as the tendency for non-branching magistral vascular distribution patterns and the lower prevalence of very high numbers of AV anastomoses in TTTS placentas may reflect a general antiangiogenic condition in TTTS placentas, in line with a recently proposed notion [35]. It is thus possible that AA anastomoses must be viewed as markers, rather than as functional determinants of TTTS.

Pathologists are occasionally asked to speculate whether the appearance of a particular diamniotic–monochorionic twin placenta may be suggestive of TTTS, especially in the context of unexplained poor pregnancy outcome. To address this question, we determined the ‘performance’ of placental examination as a post-hoc test for predicting the absence or presence of TTTS in the associated pregnancy. None of the individual placental characteristics were found to be particularly helpful. By combining certain variables, such as absence of AA anastomoses and AV imbalance, very high specificity was obtained, although the detection rate is low for this infrequent association. It needs to be emphasized that the interpretation of these data is limited by the small number of TTTS placentas and larger studies will be needed to support the validity of the proposed associations. While at present assessment of the key placental variables implicated in TTTS can only take place after delivery, it is to be foreseen that in the near future improved imaging and/or Doppler techniques will allow intrauterine detection, at which time these studies may achieve prognostic value and, possibly, assist with antenatal management of the diamniotic–monochorionic pregnancy.

As explained above, about one-third of the non-lasered TTTS group consisted of stage I TTTS, whereas all laser-treated cases had persistent Quintero stage II TTTS and above. While this indicates an overrepresentation of mild TTTS cases in the study group, we would like to emphasize that the term ‘mild’ TTTS for stage I disease is only relative, as we had several clinical stage I cases that resulted in dual fetal demise. Limited analysis of placental variables not presumed to be altered by laser surgery (such as type of cord insertion and placental share) showed no differences between laser and non-laser-treated TTTS placentas, suggesting other placental variables (such as vascular angioarchitecture) may be similar in both groups as well. Finally, we expect that findings in our study group, which includes mild TTTS cases, may apply equally to the TTTS population in general.

In conclusion, placentas of diamniotic–monochorionic pregnancies complicated by TTTS show a significantly higher frequency of velamentous cord insertion, magistral vascular distribution, uneven placental sharing, absence of AA and presence of VV anastomoses compared with non-TTTS placentas. Regardless of their functional significance, the five placental variables identified or confirmed in this study are valuable markers of TTTS and, in our opinion, deserve inclusion in the routine pathology report of diamniotic–monochorionic twin placentas. Virtually all TTTS placentas displayed AV anastomoses, suggesting these anastomoses are essential for development of the syndrome. Contrary to popular notion, however, the AV imbalance was not larger but smaller in TTTS placentas and did not correlate with the respective donor/recipient roles of the twins. This might suggest that the main role of AV anastomoses in TTTS is not the creation of an hemodynamic imbalance but, possibly, providing intertwin conduits for vasoactive and other soluble factors. The exact pathophysiology of TTTS remains largely undetermined and is certainly more complex and multifactorial than is currently recognized.

Back to Article Outline

References 

  1. Lewi L, Van Schoubroeck D, Gratacos E, Witters I, Timmerman D, Deprest J. Monochorionic diamniotic twins: complications and management options. Curr Opin Obstet Gynecol. 2003;15:177–194
  2. Ortibus E, Lopriore E, Deprest J, Vandenbussche FP, Walther FJ, Diemert A, et al. The pregnancy and long-term neurodevelopmental outcome of monochorionic diamniotic twin gestations: a multicenter prospective cohort study from the first trimester onward. Am J Obstet Gynecol. 2009;200(494):e491–e498
  3. Huber A, Hecher K. How can we diagnose and manage twin–twin transfusion syndrome?. Best Pract Res Clin Obstet Gynaecol. 2004;18:543–556
  4. Urig MA, Clewell WH, Elliott JP. Twin–twin transfusion syndrome. Am J Obstet Gynecol. 1990;163:1522–1526
  5. De Lia JE, Cruikshank DP, Keye WR. Fetoscopic neodymium:YAG laser occlusion of placental vessels in severe twin–twin transfusion syndrome. Obstet Gynecol. 1990;75:1046–1053
  6. Bajoria R, Wigglesworth J, Fisk NM. Angioarchitecture of monochorionic placentas in relation to the twin–twin transfusion syndrome. Am J Obstet Gynecol. 1995;172:856–863
  7. Denbow ML, Cox P, Talbert D, Fisk NM. Colour Doppler energy insonation of placental vasculature in monochorionic twins: absent arterio-arterial anastomoses in association with twin-to-twin transfusion syndrome. Br J Obstet Gynaecol. 1998;105:760–765
  8. Fries MH, Goldstein RB, Kilpatrick SJ, Golbus MS, Callen PW, Filly RA. The role of velamentous cord insertion in the etiology of twin–twin transfusion syndrome. Obstet Gynecol. 1993;81:569–574
  9. Gaziano EP, De Lia JE, Kuhlmann RS. Diamnionic monochorionic twin gestations: an overview. J Matern Fetal Med. 2000;9:89–96
  10. Machin GA. Velamentous cord insertion in monochorionic twin gestation. An added risk factor. J Reprod Med. 1997;42:785–789
  11. Denbow ML, Cox P, Taylor M, Hammal DM, Fisk NM. Placental angioarchitecture in monochorionic twin pregnancies: relationship to fetal growth, fetofetal transfusion syndrome, and pregnancy outcome. Am J Obstet Gynecol. 2000;182:417–426
  12. Bermudez C, Becerra CH, Bornick PW, Allen MH, Arroyo J, Quintero RA. Placental types and twin–twin transfusion syndrome. Am J Obstet Gynecol. 2002;187:489–494
  13. Benirschke K, Kaufmann P. Pathology of the human placenta. New York: Springer-Verlag; 2000;
  14. Bajoria R. Vascular anatomy of monochorionic placenta in relation to discordant growth and amniotic fluid volume. Hum Reprod. 1998;13:2933–2940
  15. Machin G, Still K, Lalani T. Correlations of placental vascular anatomy and clinical outcomes in 69 monochorionic twin pregnancies. Am J Med Genet. 1996;61:229–236
  16. De Paepe ME, Burke S, Luks FI, Pinar H, Singer DB. Demonstration of placental vascular anatomy in monochorionic twin gestations. Pediatr Dev Pathol. 2002;5:37–44
  17. Robertson EG, Neer KJ. Placental injection studies in twin gestation. Am J Obstet Gynecol. 1983;147:170–174
  18. van Gemert MJ, Sterenborg HJ. Haemodynamic model of twin–twin transfusion syndrome in monochorionic twin pregnancies. Placenta. 1998;19:195–208
  19. Umur A, van Gemert MJ, Nikkels PG, Ross MG. Monochorionic twins and twin–twin transfusion syndrome: the protective role of arterio-arterial anastomoses. Placenta. 2002;23:201–209
  20. Taylor MJ, Denbow ML, Duncan KR, Overton TG, Fisk NM. Antenatal factors at diagnosis that predict outcome in twin–twin transfusion syndrome. Am J Obstet Gynecol. 2000;183:1023–1028
  21. Quintero RA, Morales WJ, Allen MH, Bornick PW, Johnson PK, Kruger M. Staging of twin–twin transfusion syndrome. J Perinatol. 1999;19:550–555
  22. Pinar H, Sung CJ, Oyer CE, Singer DB. Reference values for singleton and twin placental weights. Pediatr Pathol Lab Med. 1996;16:901–907
  23. De Paepe ME, DeKoninck P, Friedman RM. Vascular distribution patterns in monochorionic twin placentas. Placenta. 2005;26:471–475
  24. De Paepe ME, Friedman RM, Poch M, Hansen K, Carr SR, Luks FI. Placental findings after laser ablation of communicating vessels in twin-to-twin transfusion syndrome (TTTS). Pediatr Dev Pathol. 2004;7:159–165
  25. Saunders NJ, Snijders RJ, Nicolaides KH. Therapeutic amniocentesis in twin–twin transfusion syndrome appearing in the second trimester of pregnancy. Am J Obstet Gynecol. 1992;166:820–824
  26. Bruner JP, Anderson TL, Rosemond RL. Placental pathophysiology of the twin oligohydramnios–polyhydramnios sequence and the twin–twin transfusion syndrome. Placenta. 1998;19:81–86
  27. Lopriore E, Sueters M, Middeldorp JM, Oepkes D, Walther FJ, Vandenbussche FP. Velamentous cord insertion and unequal placental territories in monochorionic twins with and without twin-to-twin-transfusion syndrome. Am J Obstet Gynecol. 2007;196(159):e151–e155
  28. Diehl W, Hecher K, Zikulnig L, Vetter M, Hackeloer BJ. Placental vascular anastomoses visualized during fetoscopic laser surgery in severe mid-trimester twin–twin transfusion syndrome. Placenta. 2001;22:876–881
  29. Machin GA, Feldstein VA, van Gemert MJ, Keith LG, Hecher K. Doppler sonographic demonstration of arterio-venous anastomosis in monochorionic twin gestation. Ultrasound Obstet Gynecol. 2000;16:214–217
  30. Bajoria R, Ward S, Chatterjee R. Brain natriuretic peptide and endothelin-1 in the pathogenesis of polyhydramnios–oligohydramnios in monochorionic twins. Am J Obstet Gynecol. 2003;189:189–194
  31. Galea P, Barigye O, Wee L, Jain V, Sullivan M, Fisk NM. The placenta contributes to activation of the renin angiotensin system in twin–twin transfusion syndrome. Placenta. 2008;29:734–742
  32. Mahieu-Caputo D, Meulemans A, Martinovic J, Gubler MC, Delezoide AL, Muller F, et al. Paradoxic activation of the renin-angiotensin system in twin–twin transfusion syndrome: an explanation for cardiovascular disturbances in the recipient. Pediatr Res. 2005;58:685–688
  33. Lewi L, Jani J, Cannie M, Robyr R, Ville Y, Hecher K, et al. Intertwin anastomoses in monochorionic placentas after fetoscopic laser coagulation for twin-to-twin transfusion syndrome: is there more than meets the eye?. Am J Obstet Gynecol. 2006;194:790–795
  34. van den Wijngaard JP, Lopriore E, van der Salm SM, Schaap AH, Vandenbussche FP, Deruiter MC, et al. Deep-hidden anastomoses in monochorionic twin placentae are harmless. Prenat Diagn. 2007;27:233–239
  35. Kusanovic JP, Romero R, Espinoza J, Nien JK, Kim CJ, Mittal P, et al. Twin-to-twin transfusion syndrome: an antiangiogenic state?. Am J Obstet Gynecol. 2008;198(382):e381–e388

PII: S0143-4004(09)00414-7

doi:10.1016/j.placenta.2009.12.024

Placenta
Volume 31, Issue 4 , Pages 269-276, April 2010

Article Tools

* Image Usage & Resolution