1 INTRODUCTION

Trisomy 18 was first described in 1960 by Edwards et al. It is the second most common autosomal trisomy in newborns, with an estimated incidence of 1/6000 live births (Rasmussen et al., 2003; Root & Carey, 1994). The syndrome typically manifests as prenatal-onset growth deficiency and developmental disability, multiple major congenital anomalies in cardiovascular, respiratory, gastrointestinal, and skeletal, and tumor susceptibility (Cereda & Carey, 2012). Trisomy 18 has been previously referred to as a “lethal anomaly” due to high mortality rate and severe developmental disability among survivors (McGraw & Perlman, 2008). In fact, withholding or withdrawal of intensive treatment had been recommended for patients with a poor prognosis (Bos et al., 1992; McGraw & Perlman, 2008; Pyle et al., 2018).

Meanwhile, questionnaire-based natural history studies by Baty et al. (1994) and Kosho et al. (2013) demonstrated that a small proportion of children with trisomy 18 can survive for long periods and continue to learn motor and communication skills. Based on these findings, the medical approaches for trisomy 18 have been changing over the past two decades (Silberberg et al., 2020). In 2004, the Japanese guidelines proposed that a treatment decision should be in the “best interests of the child,” based on deep discussions with the patients' family (Japan Society for Neonatal Health and Development, 2004; Kosho, 2008). In 2011, Nelson et al. revealed that the number of hospitalizations for trisomy 18 significantly increased from 1036 to 1616 per year between 1997 and 2009 in the United States (Nelson et al., 2012). Around the same time, an increasing number of studies performed a variety of investigations that evaluated the effectiveness of intensive care and surgical interventions on the survival of patients with trisomy 18. These studies have focused on the effects of congenital heart defects (CHD) and esophageal atresia (EA) on survival (Graham et al., 2004; Iida et al., 2020; Kaneko et al., 2008; Kosho et al., 2006; Kosiv et al., 2017; Maeda et al., 2011; Nakai et al., 2016; Nishi et al., 2014; Peterson et al., 2017). Such recent changes in the management of patients with trisomy 18 should be shared with patients' families and health providers. However, some of these studies are small case series with relatively short follow-up (Graham et al., 2004; Iida et al., 2020; Kaneko et al., 2008; Kosho et al., 2006; Maeda et al., 2011; Nakai et al., 2016; Nishi et al., 2014), and others included only patients who lived longer with fewer comorbidities (Kosiv et al., 2017; Peterson et al., 2017). Thus, it is difficult to capture the “whole picture” of the recent changes in treatment and survival of patients with trisomy 18. Here, we conducted a retrospective cohort study in patients with trisomy 18 at a Japanese pediatric tertiary referral center and further investigated how medical treatment and survival changed over a decade.

2 MATERIALS AND METHODS

3 RESULTS

3.4 Survival analysis

The Kaplan–Meier survival curves for each group are shown in Figure 1. The median survival in all the patients was 314 days (95% CI: 202–549) (Figure 1a). The median survival in the EP and LP groups was 180 days (95% CI: 56–453) and 535 days (95% CI: 299–1771), respectively (Figure 1b). The log-rank test demonstrated significantly improved survival in the LP group (p = 0.005, log-rank test).

As shown in Table 3, the overall survival rates at 1-, 2-, and 3-year was 46.4%, 35.7%, and 28.6%, respectively. In comparison to the respective survival rates, those at 1 and 2 years were not significantly different, whereas at 3 years, survival rates were significantly higher in the LP group (EP group: 13.8% vs. LP group: 44.4%, p = 0.017).

TABLE 3. Survivals in patients with trisomy 18 between the two groups

	All	EP group	LP group	p Value
	n = 56	n = 29	n = 27
1-Year survival rate	46.4 (33.9–59.4)	34.5 (19.7–56.4)	59.3 (40.3–75.8)	0.107
2-Year survival rate	35.7 (22.8–47.2)	24.1 (12.0–42.7)	48.2 (30.4–66.4)	0.094
3-Year survival rate	28.6 (18.3–41.7)	13.8 (5.3–31.5)	44.4 (27.2–63.1)	0.017

• Note: Data are presented as % (95% confidence interval). Abbreviations: EP, early period; LP, late period.

3.5 Survival to discharge

Thirty patients (52.6%) survived to discharge. Trends in survival to discharge by birth year are shown in Figure 2. The survival to discharge rate significantly increased from 27.6% (8/29) in the EP group to 81.5% (22/27) in the LP group (p < 0.001). However, medical therapies at discharge, including feeding and respiratory support, had not changed significantly between the two groups as shown in Table S3.

Among the 56 patients, 44 (78.6%) died during the study period (26 in the EP group and 18 in the LP group). The causes of death are summarized in Table 4. Chronic heart failure (CHF) was the leading cause of death in both groups. Among the 16 patients who died of CHF, 8 in the EP group and 3 in the LP group did not undergo cardiac surgery. The second leading cause of death was severe birth asphyxia in the EP group and infection in the LP group. Nine patients (three in the EP group and six in the LP group) died suddenly and unexpectedly at home (cause unknown).

TABLE 4. Causes of death in patients with trisomy 18 between the two groups

	All	EP group	LP group
	n = 44	n = 26	n = 18
Chronic heart failure	16 (36)	11 (42)	5 (28)
Severe birth asphyxia	4 (9)	4 (15)	0 (0)
Ventricular arrhythmias	3 (7)	2 (8)	1 (6)
Hepatoblastoma	3 (7)	2 (8)	1 (6)
Liver failure	2 (5)	2 (8)	0 (0)
Infection	4 (9)	1 (4)	3 (17)
Gastric perforation	1 (2)	1 (4)	0 (0)
Seizure	1 (2)	0 (0)	1 (6)
Chylothorax	1 (2)	0 (0)	1 (6)
Sudden death, cause unknown	9 (20)	3 (12)	6 (33)

• Note: Data are presented as n (%). Abbreviations: EP, early period; LP, late period.

4 DISCUSSION

In this retrospective cohort study at a Japanese pediatric tertiary referral center, the 1-year survival rate in patients with trisomy 18 increased from 34.5% during 2008–2012 (EP group) to 59.3% during 2013–2017 (LP group) (Table 3). Population-based studies of survival among patients with trisomy 18 in the United States (1999–2007) (Meyer et al., 2016) and Canada (1991–2012) (Nelson et al., 2016) reported 1-year survival rates of 13.4% and 12.6%, respectively. In Japan, a multicenter study at nine institutions affiliated with Nagoya University showed a 1-year survival rate of 29.1% among 117 newborn patients with trisomy 18 who were admitted between 2000 and 2015 (Kato et al., 2019). Compared to these prior reports, the 1-year survival rate shown in our cohort was relatively higher and increased over time. Moreover, survival rates in the LP group were still high, even at 2 and 3 years of age (48.2% and 44.4%, respectively). Although these results were obtained from a single institution study, we have shown markedly improved survival in patients with trisomy 18 in recent years.

Improved survival among patients with trisomy 18 in our study could be explained by the recent changes in surgical treatments, and this is consistent with the findings of other studies on the effects of surgeries (Graham et al., 2004; Iida et al., 2020; Kaneko et al., 2008; Kosho et al., 2006; Kosiv et al., 2017; Maeda et al., 2011; Nakai et al., 2016; Nishi et al., 2014; Peterson et al., 2017). In comparison to medical treatments in our cohort, a comparable proportion of patients in both the EP and the LP groups received intensive care, such as intubation at birth and ventilatory support. This indicates that no significant changes occurred in the initial medical management of patients with trisomy 18. In contrast, the proportion of patients who underwent surgeries significantly increased from 59% in the EP group to 96% in the LP group (Table 2). Furthermore, the postnatal ages at the first surgery were earlier in the LP group, especially for EA and CHD. An increase in the number of patients who received surgical treatment at an earlier age likely contributed to the improvement in survival over time. We speculate that this change was based on the accumulated surgical experience at our institution. Specifically, EA and CHD in trisomy 18 became gradually known not only to neonatologists, but also to surgeons and anesthesiologists. Wider recognition and better understanding of these surgical diseases might have enabled earlier decisions of surgical treatments, further led to improvement in survival.

In our study, the proportion of patients with trisomy 18 who underwent EA surgery did not significantly change over time. However, the median postnatal age during the first surgery for EA was significantly lower in the LP group (EP group: 2 days vs. LP group: 0 days), and this may have contributed to improved survival. Particularly in cases of EA with TEF, earlier surgery is potentially beneficial in order to prevent complications, such as gastric distension, subsequent gastric perforation, and respiratory failure due to leakage of the ventilatory gases through the TEF (Filston et al., 1984).

Our study clearly demonstrated that the proportion of patients with trisomy 18 who underwent cardiac surgery significantly increased from 36% to 67% over time. Furthermore, the median age at the first surgery for CHD was earlier in the LP group (EP group: 45 days vs. LP group: 28 days), as well as surgery for EA, although the difference was not statistically significant. In 2017, Peterson et al. (2017) reported that patients with trisomy 18 who died in-hospital following cardiac surgery had a high mean pulmonary artery pressure and pulmonary vascular resistance during preoperative cardiac catheterization. This indicates that cardiac surgery before the development of PH may decrease postoperative mortality in patients with trisomy 18. Recently, Tahara et al. (2021) described the histological findings of the lung tissue at the first cardiac surgery (mean age of 37 days) in patients with trisomy 18, indicating the possibility of irreversible pulmonary vascular changes due to PH, such as intimal fibrous proliferation, at an earlier age. Moreover, they suggested that abnormal lung development (medial defects and/or hypoplasia of the small pulmonary artery, alveolar hypoplasia) in patients with trisomy 18 may be associated with the earlier development of PH. Considering these findings, performing surgery earlier in those with CHD in the LP group might have contributed to improved survival.

So far, there are only a limited number of studies on the survival to discharge rate in patients with trisomy 18, and this ranges from 21% to 55% (Iida et al., 2020; Kosho et al., 2006; Nakai et al., 2016). Our study showed an overall survival to discharge rate of 52.6%. Notably, the rate significantly increased from 27.6% in the EP group to 81.5% in the LP group. Information about the increasing probability of discharging a patient home is important to parents. A questionnaire study by Janvier et al. (2016) showed that more than half of the parents of patients with trisomy 18 hoped to bring their child home and be a family. Meanwhile, they also hoped that their child would not suffer and be harmed by interventions. Our findings of increased survival to discharge rate could also facilitate better decision-making between parents and health providers.

Most children with trisomy 18 have complex chronic conditions. They are often discharged from the NICU with medical devices, including tracheostomy, home mechanical ventilation, and tube feeding. These special health care needs lead to an increased risk of emergency department visits and readmission. Therefore, care coordination should be established among the emergency department, primary care, and multiple subspecialty providers, and this will contribute to improved survival to discharge.

In our cohort, CHF remained the leading cause of death during the study period, which is in agreement with the results of other studies (Cereda & Carey, 2012; Iida et al., 2020). The CHF should be related to the progression of PH, although not confirmed with echocardiography and/or catheterization study. Palliative surgery for CHD has been reported to be effective in preventing the development of heart failure and pulmonary hypertension even in patients with trisomy 18 (Graham et al., 2004; Kaneko et al., 2008; Kosiv et al., 2017; Maeda et al., 2011; Nakai et al., 2016). Although these reports showed improved short-term survival rates, the long-term effects on survival have been controversial. In contrast, a recent case series demonstrated longer survival (> 5 years) in patients with CHD that underwent complete repair (Nakai et al., 2021; Peterson et al., 2017), suggesting that complete repair is preferable to palliative surgery for selected patients with greater weight and simple cardiac defects. Only palliative surgery was performed in most of our patients, while complete repair of ventricular septal defect was performed in only two cases. We believe that palliative surgery, such as PDA ligation and/or PAB, is insufficient to reduce long-term mortality due to CHD.

This study has some limitations. First, our findings cannot be generalized to other NICU settings because this was a single-center study. However, it allowed the evaluation of not only the overall survival but also the changes in medical and surgical treatments in detail at a single pediatric tertiary referral center. Second, due to the retrospective study design, we cannot conclusively demonstrate the relationship between the observed improved survival and changes in treatments in our cohort. Unmeasured confounding factors may underlie the observed differences between the two groups. Finally, family background, such as parental age, socioeconomic status, and family structure, were not examined in this study, and these may have influenced the survival to discharge rate and home medical care.

5 CONCLUSION

In our retrospective cohort study, the 1-year survival rate of patients with trisomy 18 markedly increased from 34.5% during 2008–2012 to 59.3% during 2013–2017. In addition, the survival to discharge rate also significantly increased from 27.6% to 81.5%. An increase in the number of patients receiving surgical treatments for EA and CHD at an earlier age may have contributed to improved survival in our cohort. The study's findings are important and may facilitate decision-making between patients' families and healthcare providers.

CONFLICT OF INTERESTS

The authors declare no conflict of interests.

AUTHOR CONTRIBUTIONS

Conceptualized the research project and drafted the initial manuscript: Shoko Tamaki and Sota Iwatani. Contributed to the acquisition of the clinical data: Ayako Izumi, Kentaro Hirayama, Dai Kataoka, Shohei Ohyama, Toshihiko Ikuta, Emiko Takeoka, Sachiko Matsui, Hitomi Mimura, Shogo Minamikawa, and Yasuo Nakagishi. Coordinated and supervised data collection, and critically reviewed the manuscript: Seiji Yoshimoto and Hideto Nakao.