Comparison of open fetal, fetoscopic and postnatal surgical repair for open spina bifida: decision analysis

ABSTRACT

Objective

Currently, there are several surgical approaches to manage fetal open spina bifida (OSB), namely postnatal surgical repair, open fetal surgery and its minimally invasive alternative, fetoscopic repair. Our objective was to determine the optimal surgical approach for OSB, weighing the benefits and risks to the fetus and the pregnant woman.

Methods

We assessed the health outcomes of open fetal, fetoscopic and postnatal surgical repair for pregnant women (mean ± SD age, 31 ± 3 years) with a singleton pregnancy and their offspring with OSB using a decision analytic model. We projected expected quality‐adjusted life years (QALYs) associated with each of the interventions, discounted at 1.5% annually over the lifetime time horizon for pregnant women and their offspring. Secondary maternal outcomes during the pregnancy included delivery mode and complications such as chorioamnionitis, uterine dehiscence, placental abruption, pulmonary embolism and death. Offspring outcomes included preterm birth, perinatal and postnatal mortality, cerebrospinal fluid (CSF) diversion surgery by 12 months of age and wheelchair use at 30 months of age. Our model was populated using data from the published literature and by consultation with clinical experts. Deterministic and probabilistic sensitivity analyses were conducted.

Results

Fetoscopic and open fetal surgery resulted in an identical number of expected QALYs (38.02 per mother–offspring dyad) and translated into a QALY gain of 1.70 per dyad compared with postnatal repair. With respect to QALYs gained, the probabilistic analyses showed that fetoscopic surgery was the preferred strategy in 51% of simulations, and open fetal surgery in the remaining 49% of simulations. When compared with postnatal repair, both open fetal and fetoscopic surgery showed that the gains in QALYs were most sensitive to the disutility associated with CSF diversion surgery and to the rate of wheelchair use. When comparing open fetal and fetoscopic approaches, the results were highly sensitive to the accuracy of all treatment effect estimates.

Conclusion

In the management of fetal OSB, both fetoscopic and open fetal surgery demonstrate superior QALY gains compared with postnatal repair, largely related to a reduced number of individuals who use a wheelchair or require CSF diversion surgery. Given similar effectiveness of fetoscopic and open fetal surgery, an individual risk assessment is essential to guide decision‐making between these two surgical approaches. © 2025 The Author(s). Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

INTRODUCTION

Open spina bifida (OSB) is a severe congenital defect affecting the central nervous system, with an incidence of 2–6 cases per 10 000 live births ¹ . This condition results from incomplete closure of the neural tube in the first month of embryonic development. The defect in the vertebrae and surrounding tissues exposes the neural placode to direct mechanical and chemical damage from the intrauterine environment. Furthermore, cerebrospinal fluid (CSF) leakage also causes hindbrain herniation and subsequent hydrocephalus. As a result, individuals with OSB experience significant lifelong impairment of spinal cord function, including sensorimotor problems in the lower limbs, bowel and bladder dysfunction, and the need for CSF diversion surgery to treat hydrocephalus ^{2 , 3} . Estimated life expectancy varies from 26–56 years depending on the level of the spinal lesion and the necessity for CSF diversion surgery ^{4 , 5} .

Since the publication of the Management of Myelomeningocele Study (MOMS) trial, open fetal surgery has become a state‐of‐the art option for the treatment of isolated OSB ⁶ . Fetal surgery offers substantial advantages compared with postnatal surgery, including a reduced need for CSF diversion surgery and enhanced lower limb motor function, ultimately improving the ability of affected toddlers to walk independently ^{6 , 7 , 8} . However, unlike postnatal surgery, the open fetal surgical technique entails significant obstetric morbidity risks, particularly the risk of preterm birth and the potential for uterine scar dehiscence or rupture, in both the index pregnancy and subsequent pregnancies, owing to the large hysterotomy required for the procedure ⁹ . In response to these challenges, alternative surgical approaches have been developed to mitigate risks during the pregnancy while preserving fetal benefits. These alternatives include minihysterotomy and fetoscopic approaches. Currently, hybrid and percutaneous fetoscopic techniques have shown promise in maintaining myometrial integrity, with no reported incidences of uterine rupture or dehiscence, even among patients undergoing vaginal delivery ¹⁰ . Consequently, a range of management strategies with differing trade‐offs for the pregnant woman and the fetus are now available for pregnancy complicated by OSB. Hence, a structured framework to systematically analyze and compare these diverse treatment options is needed.

The objective of this study was to determine the optimal surgical approach for OSB by objectively weighing the benefits and risks to the fetus and pregnant woman associated with open fetal surgery, fetoscopic surgery and postnatal surgery, using decision analysis.

METHODS

We followed the relevant non‐cost aspects of the Consolidated Health Economic Evaluation Reporting Standards and Canadian Drug Agency (CDA) guidelines for reporting this study ^{11 , 12} .

Population

We modeled a population with characteristics similar to those undergoing OSB surgery in the MOMS trial ⁶ , namely, singleton pregnancy, upper boundary of the anatomic lesion between T1 and S1 in the fetus, evidence of hindbrain herniation, normal karyotype, and absence of severe kyphosis and fetal anomalies unrelated to OSB. In the MOMS trial, the mean ± SD age of pregnant people at evaluation was 31 ± 3 years. We did not include termination of pregnancy data, as they do not impact the outcomes of OSB.

Strategies

We compared the following surgical strategies:

1. Open fetal surgery performed between 19 and 28 weeks' gestation, which included classic and minihysterectomy approaches.

2. Fetoscopic surgery performed between 19 and 28 weeks' gestation, which included hybrid and percutaneous approaches.

3. Postnatal surgery, in which a neonate undergoes surgical repair of OSB within 48–72 h after birth and recovers in the neonatal intensive care unit.

Outcomes

We reported mother–offspring dyad health outcomes measured in life years and quality‐adjusted life years (QALYs) for each strategy. A QALY equates to 1 year in perfect health. It is calculated as the product of utility weights that measures quality of life (anchored at 0, death and 1, perfect health) and time lived in a specific health state. Secondary outcomes included pregnancy‐related complications (chorioamnionitis, placental abruption, uterine dehiscence, pulmonary embolism, maternal death), delivery mode (vaginal, Cesarean section), preterm birth < 28 weeks' gestation, perinatal mortality (fetal demise or neonatal death within 28 days after delivery), postnatal mortality (between 29 days and 12 months of age, between > 12 and 30 months of age), CSF diversion surgery by 12 months of age and wheelchair use at 30 months of age. We used a lifetime time horizon for pregnant women and offspring. Mother–offspring dyad QALYs accrued over a lifetime time horizon were discounted at 1.5% annually, as recommended by the CDA guidelines ¹² .

Model structure and assumptions

We developed a decision analytic model using TreeAgePro 2020 (TreeAge Software, Williamstown, MA, USA) to project the health outcomes of open fetal surgery, fetoscopic surgery and postnatal surgery for pregnant women and their offspring with OSB (Figure 1). A hypothetical pregnant woman and offspring can undergo one of the proposed surgical interventions and experience health outcomes based on the probabilities associated with each type of surgery. The health benefits associated with the interventions were modeled through decreased necessity for CSF diversion surgery and its downstream effect on survival and quality of life, and reduced incidence of wheelchair use. The model structure accounted for intervention/delivery complications during the pregnancy and offspring outcomes, as outlined above.

Figure 1.

Decision tree comparing open fetal, fetoscopic and postnatal surgery for open spina bifida (OSB) repair. Pregnancy‐related complications include chorioamnionitis, placental abruption, uterine dehiscence, pulmonary embolism and maternal death. Transition to death for offspring is not shown; offspring may experience perinatal death (fetal death or death within 28 days after birth) or death occurring between 29 days and 12 months of age, > 12 to 30 months of age or at any time after 30 months of age. Subtree clones, with the same structure but varying probabilities, are indicated (+). CSF, cerebrospinal fluid.

We made the following assumptions about the course of the pregnancy based on surgery type: (i) following open fetal surgery, all pregnant patients were modeled to deliver by Cesarean section; (ii) patients with placental abruption were assumed to undergo emergency Cesarean section; (iii) patients who experienced pulmonary embolism had an increased risk of mortality; and (iv) in the case of maternal death, which could occur during or after pregnancy, the fetus was assumed to survive in 50% of cases. We varied the latter assumption from 0% to 100% in the sensitivity analyses.

Data

Our model was informed by data from the published literature and consultation with clinical experts (Y.K., J.W.S., K.R., A.V.K., T.V.M.). For all strategies, complications related to interventions, delivery mode, timing of delivery and neonatal and postnatal outcomes, such as the need for CSF diversion surgery, incidence of wheelchair use and mortality estimates, were based on our recently published meta‐analysis of observational and clinical studies (Table 1) ¹³ . The majority of these studies originated from high‐resource countries (57% North America, 25% Europe).

Table 1.

Key input parameters used in base‐case analysis of open fetal, fetoscopic and postnatal surgical repair for open spina bifida

Input parameter	Base case	Range	Reference
Open fetal surgery
Chorioamnionitis	0.055	0.037–0.080	13
Placental abruption	0.046	0.033–0.065	13
Pulmonary embolism	0.019	0.008–0.045	13
Uterine dehiscence	0.066	0.052–0.084	13
Vaginal delivery	0.000	0.000	13
Preterm birth (< 28 weeks' gestation)	0.045	0.019–0.099	13
CSF diversion surgery by 12 months of age	0.338	0.291–0.387	13
Wheelchair use at 30 months of age	0.290	0.185–0.408	13
Death, perinatal	0.045	0.035–0.057	13
Death between 29 days and 12 months of age	0.020	0.012–0.030	13
Death between > 12 and 30 months of age	0.044	0.012–0.095	13
Fetoscopic surgery
Chorioamnionitis	0.049	0.025–0.091	13
Placental abruption	0.063	0.014–0.245	13
Pulmonary embolism	0.003	0.001–0.021	13
Uterine dehiscence	0.000	0.000–0.010	13
Vaginal delivery	0.453	0.370–0.538	13
Preterm birth (< 28 weeks' gestation)	0.037	0.015–0.086	13
CSF diversion surgery by 12 months of age	0.430	0.375–0.486	13
Wheelchair use at 30 months of age	0.275	0.135–0.442	13
Death, perinatal	0.056	0.035–0.090	13
Death between 29 days and 12 months of age	0.022	0.010–0.050	13
Death between > 12 and 30 months of age	0.009	0.000–0.070	13, assumption
Postnatal surgery
Chorioamnionitis	0.000	0.000–0.005	13
Placental abruption	0.000	0.000–0.004	13
Pulmonary embolism, vaginal delivery	5.4/10 000	2–15/10 000	15
Pulmonary embolism, Cesarean section (OR)	2.87	2.33–3.54	17
Uterine dehiscence	0.000	0.000–0.010	13
Vaginal delivery	0.296	0.202–0.392	13
Preterm birth (< 28 weeks' gestation)	0.004	0.000–0.021	9, 44
CSF diversion surgery by 12 months of age	0.814	0.752–0.869	13
Wheelchair use at 30 months of age	0.435	0.188–0.701	13
Death, perinatal	0.012	0.001–0.040	13
Death between 29 days and 12 months of age	0.029	0.015–0.055	13
Death between > 12 and 30 months of age	0.031	0.016–0.061	13
Mortality (conditional on delivery mode and/or complication)
Pregnancy‐related death
Vaginal delivery	8/100 000	2–22/100 000	14
Cesarean section (RR)	4.50	3.40–5.80	16
Pulmonary embolism	0.012	0.006–0.019	18
Uterine rupture (OR)	4.45	1.15–17.26	19
Perinatal death
Uterine rupture (OR)	33.34	21.59–51.51	19
Placental abruption (HR)	14.80	10.90–20.00	21
Preterm birth	0.353	0.331–0.374	20
Long‐term survival (in years)
Life expectancy, individuals without CSF shunt	55.34	50–60	4
Decline in life expectancy for individuals with vs without CSF shunt	6.35	0–10	4, 23, 24, 25
Life expectancy, female at age 32 years	52.64	47–58	26

The ranges were used in the probabilistic analysis to determine parameters of input distributions.

Probabilities were sampled from beta distributions, life expectancies from gamma distributions and odds ratios (OR), hazard ratios (HR) and relative risks (RR) from log‐normal distributions.

CSF, cerebrospinal fluid.

Since the studies report zero maternal mortality (for all strategies) and pulmonary embolism (for postnatal strategy) for pregnant women, likely due to the limited sample sizes, we modelled these events conditional on delivery mode. For patients delivering vaginally, we used rates from the general Canadian pregnant population aged 30–35 years (8/100 000 for mortality and 5.4/10 000 for pulmonary embolism) ^{14 , 15} . For individuals delivering by Cesarean section, we considered an increased likelihood for these events (relative risk of 4.5 for mortality and odds ratio of 2.9 for pulmonary embolism) ^{16 , 17} . We further explored these assumptions in the sensitivity analysis. Additionally, we factored in the increased risk of pregnancy‐related mortality associated with complications specific to the surgical strategies, pulmonary embolism ¹⁸ and uterine rupture ¹⁹ , to capture the potential mortality differences between the strategies.

Furthermore, we assumed increased perinatal mortality risk associated with preterm birth ²⁰ , as well as complications such as placental abruption ²¹ and uterine dehiscence ¹⁹ . To ensure that we did not overestimate perinatal mortality, we adjusted the mortality estimate associated with the intervention to align it with the overall mortality rate reported in the meta‐analysis ¹³ . The probabilities for the remaining outcomes were derived from the reported values in the meta‐analysis, rather than being conditional probabilities given the occurrence of the complication.

Life expectancy for individuals with OSB was estimated using up to 25 years of survival data, obtained from a meta‐analysis of population‐based studies ⁴ . Since no published data exist regarding the long‐term survival of offspring who have undergone prenatal OSB repair, we assumed no differences in life expectancy between the intervention groups beyond those directly resulting from the decreased necessity for CSF diversion surgery. For CSF diversion surgery, we used only data acquired from populations with ventriculoperitoneal shunt placement because endoscopic third ventriculostomy is a relatively new CSF diversion technique ²² . We assumed that survival of individuals with OSB who received a CSF shunt is similar to that of individuals with hydrocephalus, which was estimated from three population‐based studies ^{23 , 24 , 25} . Furthermore, we conservatively assumed that individuals with OSB who received a CSF shunt and were alive at 25 years of age have a similar life expectancy to those without a CSF shunt. We incorporated the estimated life expectancy of 56 and 48 years for individuals with OSB who had not received and who had received a CSF shunt, respectively, into our decision‐analytic model. In sensitivity analyses, we varied the difference in life expectancy for individuals with OSB who had received a CSF shunt vs those who had not. We assumed that pregnant women would have a life expectancy equivalent to that of the general female population in Canada at age 32, which is an additional 52.6 years ²⁶ .

We obtained utility values for pregnant women and offspring, along with disutilities associated with procedures and related complications from the literature (Table 2). QALYs were calculated by multiplying the health‐related utility or procedure‐related disutility by the duration of each event or state.

Table 2.

Utility and disutility values for pregnant people and offspring used in base‐case analysis of open fetal, fetoscopic and postnatal surgical repair for open spina bifida (OSB)

Parameter	Base case	Range	Duration	Utility instrument	Reference
Pregnancy and postpartum period, utility
Age 25–34 years (pregnancy)	0.89	0.70–1.00	Entire pregnancy	EQ‐5D	27
Age 35–44 years	0.88	0.62–0.99	Applicable age period	EQ‐5D	27
Age 45–54 years	0.86	0.52–0.99	Applicable age period	EQ‐5D	27
Age 55–64 years	0.85	0.51–0.99	Applicable age period	EQ‐5D	27
Age 65+ years	0.86	0.60–1.00	Applicable age period	EQ‐5D	27
Intervention, disutility
Open fetal surgery	0.49	0.36–0.66	21 days	QWB	29
Fetoscopic surgery	0.25	0.61–0.90	7 days	—	Assumption
Postnatal surgery and subsequent recovery in NICU	0.13	0.50–1.00	10 days	SG	30
Pregnancy‐related complication, disutility
After vaginal delivery	0.14	0.00–0.29	12 weeks	EQ‐5D	28
After Cesarean section	0.22	0.07–0.36	12 weeks	EQ‐5D	28
After emergency Cesarean section due to placental abruption	0.24	0.09–0.38	12 weeks	EQ‐5D	28
Chorioamnionitis	0.08	0.04–0.12	4 days	Not reported	45
Pulmonary embolism	0.33	0.10–0.76	12 weeks	VAS	46
Uterine dehiscence	0.49	0.34–0.64	21 days	QWB	29
Offspring with OSB
OSB without CSF shunt, utility	0.55	0.41–0.70	Lifetime of offspring	HUI	31
OSB with CSF shunt, disutility relative to OSB without CSF shunt	0.04	0.00–0.08	Lifetime of offspring	HUI	32
Wheelchair use at 30 months, disutility relative to OSB without wheelchair use	0.10	0.04–0.16	Lifetime of offspring	HUI	31
Parent/caregiver, disutility
Parent of individual without CSF shunt	0.23	0.00–0.43	Lifetime of offspring or parent*	QWB	31
Parent of individual with CSF shunt	0.28	0.23–0.33	Lifetime of offspring or parent*	QWB	31
Parent of individual who uses wheelchair	0.28	0.23–0.33	Lifetime of offspring or parent*	QWB	31
Grieving parent, first year	0.30	0.15–0.44	1 year	TTO	33
Grieving parent, after first year	0.10	0.00–0.24	Remaining lifetime of parent	—	Assumption

The ranges were used in the probabilistic analysis to determine parameters of beta distributions, from which utilities were sampled.

^*Whichever duration is shortest.

CSF, cerebrospinal fluid; EQ‐5D, EuroQol‐5 Dimension; HUI, Health Utility Index; NICU, neonatal intensive care unit; QWB, Quality of Wellbeing Classification System; SG, standard gambling; TTO, time trade‐off; VAS, visual analog scale.

For uncomplicated pregnancy and postpartum periods, we assumed utility values to be equivalent to those of healthy females aged 25–35 years in Canada, based on the EuroQol‐5 Dimension (EQ‐5D) instrument ²⁷ . The utility decrements associated with delivery modes were also estimated based on EQ‐5D ²⁸ , while disutilities stemming from complications were obtained from the literature and were based on various assessment tools. Disutility associated with each surgical procedure (open fetal surgery, fetoscopic intervention, postnatal surgery), related complications and delivery modes (vaginal delivery, Cesarean section) was applied over the period required to resume usual activities as obtained from the literature ^{29 , 30} . Disutility associated with postnatal repair was assumed to be equal to disutility in the NICU applied for a duration of 10 days, while disutility associated with fetoscopic surgery was assumed to be half of that for open fetal surgery and was applied over 7 days.

The offspring's utility values were based on the Health Utilities Index Mark 2 (Health Utilities Inc., Dundas, ON, Canada), while disutility associated with caregiving of an offspring with OSB was based on the Quality of Wellbeing Classification System ^{31 , 32} . Utility values were lower for individuals with OSB who had received a CSF shunt and and those who required wheelchair use compared to the utility values for individuals without these conditions. In our analysis, we applied the utility of individuals with thoracic lesion OSB to represent individuals with OSB who used a wheelchair ³¹ .

Similarly, we assumed that a parent of an individual with OSB who had received a CSF shunt or used a wheelchair would have lower utility values, consistent with the utility of caregivers for individuals with a lesion at a thoracic level. Furthermore, we assumed that a parent would experience grief if their offspring were to pass away, which we approximated by equating it to disutility associated with mild‐to‐moderate depression ³³ lasting for a year and disutility of 0.1 (assumption) thereafter for a lifetime. We used an additive method to estimate utilities for joint health states.

Analyses

Two clinical experts (Y.K. and T.V.M.) reviewed the model structure for face validity. We used microsimulation to generate prognostic results for 100 000 simulated singleton pregnancies affected by OSB. Each pregnancy was followed from gestation until birth, and throughout the lifetime of both the mother and the offspring. Outcomes were reported per 100 000 pregnancies across the three surgical approaches.

In a deterministic one‐way sensitivity analysis, we varied the key model parameters over their plausible range, such as probability of CSF diversion surgery, incidence of wheelchair use, probabilities of intervention‐related complications, utility values and life expectancy of individuals with OSB, and discount rate.

We performed probabilistic sensitivity analyses to estimate the overall effect of uncertainty in our model. Transition probabilities and utility values were sampled from beta distributions, life expectancies from gamma distributions, and odds and hazard ratios from log‐normal distributions. Mean values were based on the deterministic base‐case value and variances were estimated from 95% CIs where available, or based on reasonable range otherwise (Tables 1 and 2).

RESULTS

Considering a lifetime time horizon in our simulated cohort of 100 000 pregnancies, open fetal and fetoscopic surgery resulted in an identical number of expected QALYs (38.024 and 38.023 QALYs per mother–offspring dyad, respectively), followed by postnatal repair (36.325 QALYs per dyad). Compared with postnatal repair, this translated into a gain of 1.699 and 1.698 QALYs per dyad for open fetal and fetoscopic surgery, respectively (Table 3). Open fetal and fetoscopic surgery yielded the greatest number of QALYs due to a substantial reduction in the incidence of wheelchair use and the need for CSF diversion surgery, and therefore its downstream effect on quality of life, compared with postnatal repair.

Table 3.

Pregnant person and offspring outcomes analyzed in simulated cohort of 100 000 pregnancies which underwent open fetal, fetoscopic or postnatal surgical repair for open spina bifida

Outcome	Open fetal surgery	Fetoscopic surgery	Postnatal surgical repair
QALY per mother–offspring dyad*	38.024 (33.152–42.203)	38.023 (32.859–42.506)	36.325 (31.127–40.946)
Difference in QALYs†	1.699 (0.113–3.452)	1.698 (0.209–3.316)	Ref
Probability of being optimal strategy, based on QALYs	0.485	0.512	0.003
LY per mother–offspring dyad*	65.355 (63.124–66.885)	65.628 (62.821– 67.312)	64.764 (60.169–67.381)
Difference in LYs†	0.591 (−1.760 to 3.504)	0.864 (−1.497 to 3.341)	Ref
Probability of being optimal strategy, based on LYs	0.324	0.536	0.140
Pregnant woman
Placental abruption (per 100 000 pregnancies)	4586	6365	0
Uterine dehiscence (per 100 000 pregnancies)	6626	0	0
Chorioamnionitis (per 100 000 pregnancies)	5533	5049	0
Pulmonary embolism (per 100 000 pregnancies)	1875	271	112
Vaginal delivery (per 100 000 pregnancies)	0	45 295	29 559
Death (per 100 000 pregnancies)	62	22	23
Offspring
Preterm birth (< 28 weeks' gestation) (per 100 000 pregnancies)	4566	3729	416
CSF diversion surgery (per 100 000 pregnancies)	31 650	39 688	78 099
Wheelchair use at 30 months of age (per 100 000 pregnancies)	26 019	25 212	40 378
Death, perinatal (per 100 000 pregnancies)	4540	5655	1150
Death between 29 days and 12 months of age (per 100 000 pregnancies)	1880	2036	2840
Death between > 12 and 30 months of age (per 100 000 pregnancies)	4073	837	3009

Data are given as mean (95% credibility interval (CrI)), unless stated otherwise.

CrI summarizes uncertainty in results of probabilistic sensitivity analysis.

^*Quality‐adjusted life years (QALYs) and life years (LYs) are mother and offspring years cumulatively per dyad.

^†Between open fetal and postnatal surgery and between fetoscopic and postnatal surgery.

CSF, cerebrospinal fluid; Ref, reference.

The expected number of life years, however, was the highest for fetoscopic surgery (65.63 life years per mother–offspring dyad), followed by open fetal surgery and postnatal repair (65.36 and 64.76 life years per dyad, respectively), due to fewer maternal and postnatal (> 12 to 30 months) deaths compared with open fetal surgery and a lower need for CSF diversion surgery, and therefore its downstream effect on survival, compared with postnatal repair.

In a one‐way sensitivity analysis, when comparing fetoscopic surgery and postnatal repair (Figure 2a), the disutility associated with CSF shunt, incidence of wheelchair use, the necessity for CSF diversion surgery and the difference in life expectancy between individuals who did and did not undergo CSF diversion surgery had the biggest impact on QALY gain. When comparing fetoscopic and open fetal surgery (Figure 2b), as the intervention‐related adverse events and the rates of CSF diversion surgery and wheelchair use were similar for these two strategies, a small change in treatment effect estimates or adverse event rates shifted the preference between the two approaches.

Figure 2.

Tornado diagrams of one‐way sensitivity analyses showing quality‐adjusted life years (QALYs) gained for (a) fetoscopic surgery (FS) vs postnatal repair and (b) FS vs open fetal surgery. Parameters are listed in descending order of their impact on QALY gain. Values in parentheses are range of input values for each parameter. Bars and corresponding data labels indicate QALY gain associated with maximum () and minimum () inputs for each parameter. CSF, cerebrospinal fluid; d, days; LE, life expectancy; m, months.

The probabilistic analyses showed that, with respect to QALYs gained, fetoscopic surgery was the preferred strategy in 51% of simulations and open fetal surgery in 49% of simulations, while with respect to life years gained, fetoscopic surgery was the preferred option in 54% of simulations, open fetal surgery in 32% of simulations and postnatal repair in the remaining 14% (Table 3).

DISCUSSION

We evaluated the optimal surgical approach for managing a singleton pregnancy complicated by OSB over the lifetime time horizon of a mother–offspring dyad. Our findings showed that fetoscopic and open fetal surgery outperform postnatal repair in terms of QALYs gained, both resulting in nearly identical QALYs. Conversely, since maternal and postnatal (> 12 to 30 months) mortality was higher with open fetal surgery, fetoscopic surgery resulted in a slightly higher number of total life years gained. Differences were, however, fairly small compared with postnatal surgery (1.7 QALYs and 0.9 life years), questioning their clinical relevance.

Fetoscopic and open fetal surgery for OSB emerged as the interventions with the highest QALYs for both the mother and the affected child, primarily due to a substantial reduction in the requirement for CSF diversion surgery and a lower incidence of wheelchair use, both of which were associated with substantial disutility as well as shorter life expectancy. These advantages translate to higher QALYs for both the mother and the affected child compared with a postnatal repair approach. Conversely, complications during pregnancy had a relatively small effect on the results due to their low incidence rates and short duration, which inherently limited their disutility to the mother–offspring dyad. Fetoscopic surgery for OSB resulted in a slightly higher number of life years for the dyad due to fewer offspring deaths between > 12 and 30 months of age and fewer complications during the pregnancy, such as uterine dehiscence/rupture and pulmonary embolism, which contribute directly to mortality for the pregnant woman. Furthermore, the risk of uterine dehiscence or rupture, an inherent risk from any uterine incisional scar, may extend into subsequent pregnancies. Recent publications indicate an approximate 10% incidence of uterine rupture in subsequent pregnancies following open fetal surgery for OSB ^{34 , 35 , 36} .

Since the publication of the MOMS trial ⁶ , the postnatal advantages of open fetal surgery for OSB, notably in reducing the need for CSF diversion surgery, have become clear, and this benefit continues into the school‐age years ⁸ . More recent publications show a comparable rate of CSF diversion surgery by 12 months and wheelchair use at 30 months of age following fetoscopic surgery when compared with open fetal surgery for OSB ^{37 , 38 , 39 , 40 , 41} .

We expand on previous work by concurrently comparing three surgical approaches. Two previous studies employing modeling techniques compared postnatal repair with open fetal surgery or open fetal surgery with fetoscopic surgery. Werner et al. ⁴² assessed the cost‐effectiveness of postnatal and open fetal surgery. Consistent with our results, they observed that open fetal surgery was associated with a higher number of QALYs compared with postnatal repair. Packer et al. ⁴³ assessed the cost‐effectiveness of open fetal vs fetoscopic surgery, showing that fetoscopic surgery is substantially more effective in terms of QALYs gained, whereas our analysis found nearly identical QALY outcomes between the two approaches. The discrepancy between the findings may be explained by differences in how we quantified the effectiveness of the interventions. We conservatively modeled the health benefits associated with the interventions through reduced incidence of wheelchair use, decreased necessity for CSF diversion surgery and the downstream effect on survival and quality of life. Packer et al. ⁴³ quantified the benefits through better/worse motor response levels relative to the anatomic lesion, which may require additional assumptions when it comes to associated life expectancy and utility/disutility. Furthermore, they did not include CSF diversion surgery as an outcome, which might impact the results substantially, as demonstrated in our deterministic sensitivity analysis.

This study has several limitations. First, to date, there are no studies comparing directly the effectiveness and safety of all three interventions. Thus, inputs for effectiveness and complications were estimated separately, based on meta‐analyses. Given the small difference in QALY gains associated with open fetal and fetoscopic surgery, the results are likely to be influenced by the accuracy of these treatment effect estimates. While our probabilistic analyses addressed the variability in estimates, the relative uncertainty in long‐term outcomes between the two procedures should be considered when interpreting our findings. Second, we did not account for the impact of the interventions on future pregnancies due to lack of published literature on subsequent pregnancies after fetoscopic repair for OSB. Complications such as uterine rupture and dehiscence during the index pregnancy, associated with open fetal surgery, may reduce the QALYs gained for these surgeries compared with postnatal and fetoscopic surgical repair, if subsequent pregnancies are considered. Third, there is a lack of data on life expectancy of individuals with OSB, especially considering recent advances in medicine. We conservatively assumed that individuals with OSB, with and without a CSF shunt, have similar survival after 25 years of age, which may not hold true. Many individuals remain shunt‐dependent and complications can arise later in life that would make open fetal surgery even more attractive, as shown in Figure 2b. Future studies with extended follow‐up are essential to assess comprehensively the effectiveness of prenatal interventions for OSB, including the impact on subsequent pregnancy outcomes, the long‐term need for CSF diversion surgery and potential deterioration in motor function. Finally, this study does not account for variability in surgical outcomes based on institutional or surgeon expertise. The difference in success rate and complication risks between high‐ and low‐experience centres may impact the overall effectiveness of the procedures.

Our model has several strengths. We considered the health outcomes of both the pregnant woman and the offspring over appropriate time horizons. Furthermore, we considered preferences related to combined mother–offspring health states and incorporated the disutilities associated with caregiving for individuals with OSB. However, we acknowledge that dedicated patient preference studies may provide valuable insights into the values and priorities of expectant mothers navigating decisions on OSB surgery. Additionally, our model relied predominantly on synthesized data from a very recent meta‐analysis, offering a broader perspective on the outcomes of different surgical strategies.

In conclusion, based on current evidence, for the management of fetal OSB, both open fetal surgery and fetoscopic surgery outperform postnatal repair in terms of QALYs gained for the mother–offspring dyad when considering the index pregnancy. Of note, the fetoscopic approach resulted in a slightly higher number of life years gained. Our ‘objective’ risk‐benefit analyses may not always align with the patient perspective because individuals may prioritize outcomes differently based on their values and circumstances. In addition, clinician expertise and preferences also influence the decision‐making process. The results of the present study are reassuring in this regard because the differences in both QALYs and life years between the different surgical approaches are small. This study underscores the need for a nuanced approach that balances evidence‐based guidelines with patient goals and clinical expertise, fostering personalized and effective care.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.