Results:
A total of 418 TTTS consecutive cases underwent laser surgery during the study period at both institutions. UA PI of both donor and recipient twins was available in 89.5% (374/418) of TTTS cases and DUAPI was calculated in the latter cases. DUAPI was available in 96.4% (54/56), 99.2% (132/133), 81% (158/195), and 88.2% (30/34) of Quintero stage I, II, III and IV cases, respectively. UA PI was recorded in only 65% (34/52) of cases when the UA EDF was absent or reversed in either twin, even though PI could be estimated in these cases. All cases where UA PI was available in both twins were included in the study (n=374); 6.7% (25/374) of these cases were lost to follow-up to 30 days of life, leaving 349 cases for analysis. For the analysis, one participating center contributed with 118 cases and the other one with 231 cases. TTTS Quintero stage I or II was diagnosed in 176 cases and TTTS Quintero stage III or IV was diagnosed in 173 cases. In the whole cohort, double twin survival and survival of at least one twin to 30 days was observed in 67% (234/349) and 90.3% (315/349), respectively. The demographic and clinical characteristics according to Quintero stages and DUAPI results are described in Tables 1 and 2.
ROC curve analysis demonstrated that higher inter-twin DUAPI was associated with reduced dual infant survival to 30 days of life [area under the curve (AUC): 0.67; p<0.001]. These results are consistent with an earlier report at a single institution9. Of note, ROC curve analysis demonstrated that higher UA PI in donor twin, (AUC: 0.66; p<0.001) but not in the recipient twin (AUC: 0.45; p=0.1), was also associated with reduced dual infant survival to 30 days of life. These observations suggest that the association of DUAPI with double infant survival in TTTS cases undergoing laser surgery is primarily driven by UA PI in the donor twin. We chose to use DUAPI instead of UA PI in the donor twin because it does not have a significant correlation with gestational age at ultrasonography (spearman’s rho correlation coefficient: -0.03; p=0.6). As expected, UA PI in the donor twin had a significant correlation with gestational age at ultrasonography (spearman’s rho correlation coefficient: -0.16; p=0.02). Thus, the clinical use of UA PI in the donor twin to predict infant survival in TTTS requires the use of a normogram. In contrast, intertwin DUAPI does not change with gestational age, and the use of a single cutoff (<0.4) is simple to use.
The donor twin had higher UA PI that the recipient twin in 56.7% (198/349) of TTTS cases and sFGR was identified in 21.5% (75/349) of cases. Of note, in only 66.7% (50/75) of TTTS cases with sFGR, the donor twin had higher UA PI than in the recipient twin; we anticipated this proportion to be higher because of the known association of sFGR with increased impedance to blood flow in the UA.
Significant differences in double twin survival to 30 days of life was seen between DUAPI groups in the whole cohort [<0.4: 76.8% (162/211) vs. ≥0.4: 52.2% (72/138); p<0.001] and in the subgroup of women with TTTS Quintero stage I or II combined [<0.4: 77.8% (105/135) vs. ≥0.4: 58.5% (24/41); p=0.015] as well as in women with TTTS Quintero stage III or IV combined [<0.4: 75% (57/76) vs. ≥0.4: 49.5% (48/97); p=0.001] (see Table 2). Similar findings were observed when double twin survival to 30 days of life was compared between the two DUAPI groups in TTTS cases with Quintero stage I individually [<0.4: 88.1% (37/42) vs. ≥0.4: 54.5% (6/11); p=0.02] or Quintero stage III individually [<0.4: 80.3% (49/61) vs. ≥0.4: 50% (43/86); p<0.001]. Among women with Quintero stages II or IV as individual groups, those with a DUAPI <0.4 had higher double twin survival to 30 days than those with a DUAPI ≥0.4, but these differences did not reach statistical significance in either group {Quintero stage II [<0.4: 73.1% (68/93) vs. ≥0.4: 60% (18/30); p=0.1] or Quintero stage IV [<0.4: 53.3% (8/15) vs. ≥0.4: 45.5% (5/11); p=0.6]}.
No significant differences in survival of at least one twin to 30 days of life was seen between DUAPI groups in the whole cohort [<0.4: 91% (192/211) vs. ≥0.4: 89.1% (123/138); p=0.6] or in the subgroup of women with TTTS Quintero stage I or II combined [<0.4: 91.9% (124/135) vs. ≥0.4: 90.2% (37/41); p=0.7] or in women with TTTS Quintero stage III or IV combined [<0.4: 89.5% (68/76) vs. ≥0.4: 88.7% (86/97); p=0.9]. In a multivariable regression analysis, DUAPI <0.4 was not associated with survival of at least one infant to 30 days [adjusted odds ratio (aOR): 0.64, 95% confidence interval (CI): 0.19-2.2; p=0.5] in the whole study cohort or in the subgroup analyses.
Intertwin DUAPI <0.4 was associated with a 3-fold to 3.5-fold increase in double twin survival to one month of age in the regression model of the whole cohort and the subgroup analysis (See Tables 3, 4 and 5). This was determined using DUAPI ≥0.4 as a reference when the analysis was adjusted for confounding variables. Indeed, intertwin DUAPI <0.4 was associated with increased survival of both twins to one month of age in the whole cohort (aOR: 3.40; 95% CI: 2.02-5.70; p<0.001) (Table 3), in the subgroup of women with TTTS Quintero stage III or IV (aOR: 3.23; 95% CI: 1.52-8.85; p=0.002) (Table 4) as well as in those with TTTS Quintero stage I or II (aOR: 3.05; 95% CI: 1.32-7.09; p=0.009) (Table 5). As anticipated, an additional variable associated with increased double twin survival to one month of age was gestational age delivery (See Tables 3, 4 and 5). In contrast, neither Quintero staging nor any of the sonographic criteria used in Quintero staging or sFGR were associated with infant survival when the analysis was adjusted for confounding factors (See Tables 3, 4 and 5). Similar findings were observed when the MVP of the donor (aOR: 0.91; 95% CI: 0.70-1.19; p=0.5) or recipient twin (aOR: 0.98; 95% CI: 0.89-1.10; p=0.8) were included as continuous variables in the regression models.
There were significant differences in the proportion of women with double infant survival to 30 days according to Quintero staging [81.1% (43/53), 69.9% (86/123), 62.6 (92/147) and 50% (13/26) for Quintero stage I, II, III and IV, respectively; p=0.002]. In contrast, no significant differences were observed for survival of at least one twin to 30 days according to Quintero staging [92.5% (49/53), 91.1% (112/123), 90.5 (133/147) and 80.8% (21/26) for Quintero stage I, II, III and IV, respectively; p=0.2]. The observation that Quintero staging was associated with double infant survival in the univariate analysis but not after the analysis was adjusted for DUAPI<0.4 and other confounders in the multivariable regression model (Table 3) indicates that DUAPI <0.4 supersedes the Quintero classification in the prediction of double infant survival to 30 days of life in TTTS cases.