OBJECTIVE. Our goal was to investigate the neonatal mortality rate and the mortality rate during the NICU stay for extremely low birth weight infants born in Japan in 2005.
METHODS. The Committee of Neonatal Medicine of the Japan Pediatric Society retrospectively surveyed the deaths of extremely low birth weight infants born and hospitalized between January 1 and December 31, 2005. From 297 institutions in Japan, data on 3065 extremely low birth weight infants, which represented 98.4% of those reported in the maternal and health statistics of Japan in 2005, were collected.
RESULTS. The neonatal mortality rate and the mortality rate during the NICU stay were 13.0% and 17.0%, respectively, which were lower than 17.7% and 21.5% in the survey in 2000. The neonatal mortality rates and the mortality rates during the NICU stay were 53.3% and 67.7% in the <400-g birth weight group (n = 62), 42.1% and 53.5% in the 400- to 499-g birth weight group (n = 159), 22.2% and 27.7% in the 500- to 599-g birth weight group (n = 387), 16.8% and 22.2% in the 600- to 699-g birth weight group (n = 537), 9.4% and 12.7% in the 700- to 799-g birth weight group (n = 574), 6.3% and 9.1% in the 800- to 899-g birth weight group (n = 649), and 3.9% and 5.3% in the 900- to 999-g birth weight group (n = 697), respectively. The factors involved in the deaths of extremely low birth weight infants included lower gestational age, lower birth weight, male gender, multiple birth, institutions in which <10 extremely low birth weight infants were admitted per year, and no prenatal maternal transfer.
CONCLUSION. The mortality rates of extremely low birth weight infants who were born in 2005 demonstrated definite improvement.
Advances in perinatal technology, such as pulmonary surfactant replacement therapy, high-frequency oscillatory ventilation, and prenatal administration of glucocorticosteroids, have resulted in decreasing mortality rates for preterm infants. Perinatal care systems, such as regionalization and prenatal maternal transfer, also have decreased mortality rates for infants.
The Committee of Neonatal Medicine of the Japan Pediatric Society has investigated the mortality rates for extremely low birth weight (ELBW) infants (birth weight [BW] of <1000 g) every 5 years since 1990. The subjects of the national surveys in 1990, 1995, and 2000 represent >90% of the ELBW infants from the maternal and child health statistics of Japan. Therefore, the surveys provide highly reliable data. The data represent some of the most important indices for the evaluation of perinatal and neonatal management. Families are given clear information about the serious problems of delivering extremely immature infants, including their chances of survival and possible complications. In the latest 3 surveys, the mortality rates for ELBW infants have been improving.1–3 The objectives of this study were to investigate the mortality rate of ELBW infants in 2005, to compare the data with those for previous years, and to elucidate factors influencing mortality rates.
After approval by the institutional review board of the Japan Pediatric Society, questionnaires were sent to 1199 institutions, including hospitals with >100 beds in which obstetrics and pediatrics services were established, children's medical centers, and perinatal medical centers, to investigate retrospectively ELBW infants who were born between January 1 and December 31, 2005, and were admitted to the NICU. The institutions in which ELBW infants were admitted were requested to respond with the following items: survival/death of the respective ELBW infants and postnatal age at death, immediate cause of death, gestational age, BW, gender, outborn/inborn status, prenatal maternal transfer, and multiple birth. The questionnaires were sent in July 2006 and collected at the end of December 2006. A total of 1035 institutions replied (response rate: 86%); 297 of them had admitted 3065 ELBW infants, which represented 98.4% of those reported in the maternal and child health statistics of Japan in 2005.4 ELBW infants were not admitted in 738 institutions because those institutions had level I or II nurseries or had abandoned perinatal care services. Of the 297 institutions, 125 (42.1%) admitted ≥10 ELBW infants each, for a total of 2375 (77.5%). The subjects included 1527 male infants (49.8%), 2858 inborn infants (93.3%), 2477 singleton births, and 1961 infants (64.0%) whose mothers were transferred prenatally. Gender and gestation period were unknown for 1 and 3 infants, respectively. The gestational age generally was evaluated on the basis of the last menstrual period or ultrasound measurements during early pregnancy; if needed, the Ballard score5 was used.
Mortality rates according to BW or gestational age in 2005 were compared with those in 2000 by using χ2 tests. Comparisons of the cumulative survival rates with postnatal age according to BW or gestational age were performed by using Kaplan-Meier survival analysis. To analyze the factors involved in death during the NICU stay, multivariate logistic regression analyses were conducted with the dependent variable of the existence/absence of death and objective variables including gestational age, BW, gender, delivery in institutions, prenatal maternal transfer, multiple births, and institutions in which ≥10 ELBW infants were admitted per year. The statistical analyses were conducted by using Doctor SPSS II for Windows (SPSS Japan, Tokyo, Japan). Results with P values of <.05 were determined to be significantly different.
The mortality rate for 3065 ELBW infants (<28 day of age) was 13.0%, and the mortality rate during the NICU stay was 17.0%. In comparisons according to BW, both the neonatal mortality rate and the mortality rate during hospitalization were higher in the lower-BW groups (Table 1).
For the subjects except the 3 infants whose gestational age was unknown, the neonatal mortality rates according to gestational age and the mortality rates during the NICU stay were higher for the more-premature infants (Table 2). In comparison with the group with gestational ages of ≥28 weeks, the risks of neonatal death and death during the NICU stay were significantly higher for the groups at <25 weeks of gestation. However, these risks were not significantly different in the 26- and 27-week gestational age groups.
Cumulative Mortality Rates
We collected data until NICU discharge. However, analysis of the cumulative mortality rates for BW groups or gestational age groups were limited to 90 days of postnatal life, because the hospitalization of surviving ELBW infants in our previous study was for ≥90 days after birth6 and no information about death after hospital discharge was available. The results of log-rank tests revealed a significant difference in the cumulative mortality rates according to BW (P < .0001). In comparison with the 900- to 999-g BW group, there were significant differences in the cumulative mortality rates for the <400-g BW group (P < .0001), the 400- to 499-g BW group (P < .0001), the 500- to 599-g BW group (P < .0001), the 600- to 699-g BW group (P < .0001), the 700- to 799-g BW group (P < .0001), and the 800- to 899-g BW group (P = .0078). With the exception of the <400-g BW and 400- to 499-g BW groups, there were significant differences in the cumulative mortality rates between groups. In the <700-g BW groups, mortality rates increased in the newborn period and increased gradually thereafter (Fig 1).
As shown in Fig 2, there were significant differences in the cumulative mortality rates according to gestational age (P < .0001). There was no significant difference in cumulative mortality rates between the ≥28-week gestational age group and the 26- and 27-week gestational age groups or between the 26- and 27-week gestational age groups. However, significant differences were found between other groups. For infants who were born at gestational ages of 22, 23, or 24 weeks, mortality rates increased rapidly in the newborn period and increased gradually thereafter.
Factors Involved in Death During the NICU Stay
The significant factors involved in death during the NICU stay were gestational age, BW, institutions in which ≥10 ELBW infants were admitted per year, multiple pregnancy, and prenatal maternal transfer. Delivery in/out of the institution was not a significant factor (Table 3).
Mortality Rates According to District
After adjustment for the factors of gestational age, BW, gender, multiple birth, prenatal maternal transfer, and institutions in which ≥10 ELBW infants were admitted per year, multivariate logistic regression analysis was conducted on the risk of death during the NICU stay. Compared with the NICU mortality rate in Tokyo, where the number of admissions was greatest, significantly higher risk was found in 7 prefectures among 46 districts. The highest risk ratio was 4.9 (95% confidence interval: 1.8–13.5; P = .002), compared with the mortality rate in Tokyo.
Comparison With Previous Survey Results
The neonatal mortality rate and the mortality rate during the NICU stay for ELBW infants (<800-g BW) who were born in 2005 were significantly lower than those for infants who were born in 2000 (Figs 3 and 4). The neonatal mortality rates for infants according to gestational age in 2005 were significantly lower than those in 2000 for all gestational age categories, and the mortality rates during the NICU stay were also significantly lower for the <28-week gestational age groups (Figs 5 and 6).
The results of this survey showed that the NICU mortality rates for ELBW infants born in 2005 were substantially lower than those for infants born in 2000. Unfortunately, maternal background (eg, delivery method, chorioamnionitis, premature rupture of membranes, pregnancy-induced hypertension, and prenatal administration of glucocorticosteroids) and detailed postnatal treatment were not investigated. Therefore, we could not elucidate what factors contributed to the reduction in mortality rates.
The mortality rates during the NICU stay for the <900-g BW group and the ≥25-week gestational age group of ELBW infants who were born in 2005 were significantly greater than those for the 900- to 999-g BW group and the ≥28-week gestational age group, respectively. In particular, the mortality rates for the <500-g group and the ≤23-week group were markedly higher, resulting in an increase in deaths. As indicated by the cumulative mortality data, the increase depended on a high mortality rate during the newborn period. Although the mortality rates for infants born in 2005 were improved, compared with those for infants born in 2000, the survival of ELBW infants with <500-g BWs or <23-week gestational ages remains difficult.
Surveys by the Vermont Oxford Network from 1996 to 2000 reported that the mortality rate for 4176 ELBW infants with 401- to 500-g BWs was 83%, and 2186 of those infants (52%) died in the delivery room.7 The deaths of such a large number of infants in the delivery room are thought to depend on the decisions of neonatal physicians who attended the delivery; in Japan, considering the health care economics and the differences in social and cultural backgrounds, the delivery room mortality rate is probably lower. In the survey of newborns in 2005, infants with <400-g BWs numbered 62 (28 in 2000), those with 400- to 499-g BWs numbered 159 (126 in 2000), and those with 22- and 23-week gestational ages numbered 379 (272 in 2000).3 The numbers of infants who were treated in the NICU in 2005 were clearly increased, compared with those in 2000. Therefore, the probability is low that physicians selected infants who had a high probability of survival, resulting in the reduction in mortality rates.
The independent factors involved in increased mortality rates during the NICU stay for ELBW infants were male gender and multiple births, in addition to premature gestational age and low BW. Conversely, the factors of an institution in which ≥10 ELBW infants were admitted per year and delivery after transportation of the mother to the hospital significantly reduced mortality rates. The findings of this study that male gender and multiple birth, in addition to gestational age and BW, are risk factors for the death of ELBW infants are consistent with the results of previous investigations.8,9 The mortality rate of ELBW infants who were born in hospital was significantly lower than that of infants who were born at other hospitals and transferred in.10 In studies conducted in Japan, the differences in delivery between outborn and inborn infants was not a significant mortality factor. It is assumed that neonatal physicians in children's medical centers without obstetric facilities went to an obstetric institution and attended the expected delivery of ELBW infants, subsequently transporting the newborns to the children's medical center.
The fact that infants who were admitted in an institution in which ≥10 ELBW infants were admitted per year had less risk of death, compared with infants who were admitted in an institution in which <10 infants were admitted per year, probably suggests that hospitalization of ELBW infants in specific institutions reduces mortality rates. However, even among the domestic tertiary perinatal medical centers, there are differences in mortality rates and incidences of complications.11 Therefore, the problems cannot be solved solely by admitting ELBW infants to specific institutions. The aforementioned factors might be significant but are not substantially involved in the NICU mortality rate (Nagelkerke R2 = 0.229), and other factors exert a major influence.
For death during the early period after birth, especially <12 hours after birth, it has been reported that the factors involved include nonadministration of pulmonary surfactant, endotracheal intubation in the delivery room, nonadministration of steroids to mothers, low Apgar scores 1 minute after birth, male gender, multiple pregnancy, and nonadministration of tocolytic agents, in addition to the factors of less-premature gestational age and lower BW.12 As indicated in this report, early death cases included infants who were intentionally not treated (ie, nonreplacement of pulmonary surfactants and nonexecution of positive resuscitation). In this study, the possibility that some of the infants who died during the early postnatal period corresponded to such cases cannot be ruled out. The mortality rate during the NICU stay can be influenced by withholding treatment of infants with obviously poor prognoses in the NICU or abandoning treatment.13 The guidelines for resuscitation of ELBW infants have not been unified for countries or institutions. Therefore, in comparisons of mortality rates between countries and institutions, these issues should be considered. In Japan, it is common to try to resuscitate ELBW infants who are born at 22 to 23 weeks of gestation or whose BWs are <500 g when the infants seem to be viable. The reasons for this may include a lack of information on long-term outcomes for ELBW infants with BWs of <500 g or gestational ages of <24 weeks and the medical care benefits provided by local governments for low BW infants or sick newborns. Withdrawal of care is usually performed in Japan when ELBW infants do not respond to resuscitation in the delivery room or are in extremely poor condition after prolonged treatment in the NICU.
In this study, it was shown that there were differences between districts in the mortality rates of ELBW infants during the NICU stay. This finding is important information for establishing local perinatal care systems. Regional differences in mortality rates warrant thorough audits to assess causalities.
Constant improvement in mortality rates for ELBW infants in Japan was observed, in contrast to rates in other countries.14,15 The birth rates for ELBW and very low BW infants in Japan in 2005 were 3 per 1000 live births and 8 per 1000 live births, respectively. These birth rates are lower than those in the United States in 2005.16 The reasons for this may include the improvement of socioeconomic conditions, higher levels of education for women, comprehensive health care provided by the government, regionalization for high-risk pregnancies, and medical advances in the care of pregnant women and neonates.17 The reduction of mortality rates for ELBW infants causes new problems. One of those problems is the fact that the survival of less-premature infants involves prolonged NICU stays. Even among nonsurvivors, ∼50% died after the newborn period. Another is a problem regarding morbidities and developmental outcomes. The survey by the Vermont Oxford Network reported that 36% of infants with 401- to 500-g BWs who were treated in the NICU survived. However, they frequently developed severe retinopathy of prematurity and severe chronic pulmonary disease.7 In a domestic survey in Finland, there was no major change in mortality rates between infants who were born in 1996–1997 versus 1999–2000; however, rates of complications, including respiratory distress syndrome, intraventricular hemorrhage, severe chronic lung disease, and sepsis, were increased.15 A study in Japan showed that the rate of chronic lung disease in 2000 was increased, compared with that in 1995.18 Sufficient data have not been accumulated for investigation of changes in the incidence rates of other morbidities and developmental outcomes with the improvement in mortality rates. Follow-up data at 3 years of age collected by a multicenter neonatal research network in Japan showed that the proportion of ELBW children with cerebral palsy increased from 12.0% in 1990 to 16.3% in 2000. However, 355 (29.3%) of 1208 survivors in 1990 and 994 (55.7%) of 1784 survivors in 2000 were lost to follow-up monitoring in those studies (Y. Uetani, MD, unpublished data, 2008). Reliable outcome data on the survival of ELBW infants are needed for evaluation of current perinatal care. In particular, follow-up data for ELBW infants with BWs of <500 g and those born at 22 to 23 weeks of gestation are necessary to determine whether it is reasonable to continue to try to save these smallest infants in Japan. Previous reports showed high incidences of poor neurologic and developmental outcomes.19,20
The mortality rate for ELBW infants who were born in 2005 in Japan demonstrated definite improvement. However, regional differences in mortality rates were observed. To reduce mortality rates in the future, it is important to promote prenatal maternal transfer and centralization of admission of ELBW infants, as well as analyzing causes of death. Although reduction of mortality rates is important, it is also important for family members and perinatologists to understand what changes in the outcomes of ELBW infants have occurred.
We express our sincere gratitude to the physicians of the institutions in Japan who participated in this survey. We are also grateful to the local service officials and physicians responsible in the respective regional groups. Institutions and representative physicians in this survey were as follows: T. Hattori (Sapporo Municipal Hospital), K. Amizuka (Aomori Prefectural Central Hospital), R. Goto (Akita Red Cross Hospital), M. Watanabe (Yamagata Prefectural Central Hospital), T. Sakai (Miyagi Children's Hospital), S. Chida (Iwate Medical University), J. Ujiie (Fukusima National Hospital), T. Ono, H. Kimoto (Saitama Children's Hospital), M. Tamura (Saitama Medical Center), Y. Honma (Jichi Medical University), T. Koizumi (Gunma Children's Hospital), K. Kabe (Aiiku Hospital), M. Sugiura (Kyorin University), K. Sawada (Toho University Sakura Hospital), Y. Miyamoto (Ibaraki Children's Hospital), Y. Itani (Kanagawa Children's Hospital), A. Yamazaki (Niigata Municipal Hospital), T. Futatani (Toyama University), M. Kubo (Ishikawa Prefectural Central Hospital), H. Tsukahara (Fukui University), T. Nakamura (Nagano Children's Hospital), N. Mizobe (Yamanashi Prefectural Central Hospital), M. Oshiro (Ogaki Municipal Second Hospital), S. Oki (Seirei Hamatsu Hospital), C. Suzuki (Red Cross Nagoya First Hospital), Y. Uetani (Hyogo Children's Hospital), S. Shimada (Kyoto Baptist Hospital), K. Hashimoto (Red Cross Shiga Hospital), Y. Takahashi (Nara Medical University), R. Higuchi (Wakayama Medical University), H. Yamamoto (Mie Chuou Medical Center), H. Kitajima (Osaka Medical Center and Research Institute for Maternal and Child Health), F. Kato (Shimane Prefectural Central Hospital), I. Nagata (Tottori University), S. Watabe (Kurashiki Central Hospital), M. Hayashidani (Hiroshima Municipal Hospital), K. Hasegawa (Yamaguchi University), K. Kikkawa (Kochi Medical Center), M. Kajiwara (Ehime Prefectural Central Hospital), T. Saijyo (Tokushima University), K. Isobe (Kagawa University), K. Sato (National Kyushu Medical Center), M. Aoki (National Nagasaki Medical Center), Y. Kondo (Kumamoto Municipal Hospital), K. Iida (Oita Prefectural Central Hospital), S. Ibara (Kagoshima Municipal Hospital), T. Takayanagi (National Saga Hospital), Y. Kodama (Miyazaki University), and M. Kohama (Okinawa Prefectural Chubu Hospital).
- Accepted May 19, 2008.
- Address correspondence to Kazuo Itabashi, MD, Showa University School of Medicine, Department of Pediatrics, 1-5-8 Hatanodai, Shinagawa-ku, Tokyo 142-8666, Japan. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on this Subject
With the development of prenatal and neonatal intensive care during the past 2 decades, the survival rates for ELBW infants have improved.
What This Study Adds
The mortality rate for ELBW infants who were born in Japan in 2005 demonstrated definite improvement in contrast to that in other countries.
- ↵Japan Pediatric Society, Committee of Neonatal Medicine, Subcommittee for the Survey on High Risk Newborns. Studies on the state of care for high risk neonate and neonatal mortality in our country (year 1990) [in Japanese]. J Jpn Pediatr Soc.1991;95 (11):2454– 2461
- Japan Pediatric Society, Committee of Neonatal Medicine, Subcommittee for the Survey on High Risk Newborns. Studies on the state of care for high risk neonate and neonatal mortality in our country (year 1995) [in Japanese]. J Jpn Pediatr Soc.1996;100 (12):1931– 1938
- ↵Japan Pediatric Society, Committee of Neonatal Medicine, Subcommittee for the Survey on High Risk Newborns. Studies on the state of care for high risk neonate and neonatal mortality in our country (year 2000) [in Japanese]. J Jpn Pediatr Soc.2002;106 (4):603– 613
- ↵Mothers' and Children's Health and Welfare Association. Live births and percentages by birth weight and sex, 1951–2006. In: Maternal and Child Health Statistics of Japan 2007. Tokyo, Japan: Mothers' and Children's Health and Welfare Organization; 2007:44–45
- ↵Lucey JF, Rowan CA, Shiono P, et al. Fetal infants: the fate of 4172 infants with birth weights of 401 to 500 grams: the Vermont Oxford Network experience (1996–2000). Pediatrics.2004;113 (6):1559– 1566
- ↵Kusuda S, Fujimura M, Sakuma I, et al. Morbidity and mortality of infants with very low birth weight in Japan: center variation. Pediatrics.2006;118 (4). Available at: www.pediatrics.org/cgi/content/full/118/4/e1130
- ↵Markestad T, Kaaresen PI, Ronennestad A, et al. Early death, morbidity, and need of treatment among extremely premature infants. Pediatrics.2005;115 (5):1289– 1298
- ↵Barton L, Hodgman JE. The contribution of withholding or withdrawing care to newborn mortality. Pediatrics.2005;116 (6):1487– 1491
- ↵Meadow W, Lee G, Lin K, Lantos J. Changes in mortality for extremely low birth weight infants in 1990s: implications for treatment decisions and resource use. Pediatrics.2004;113 (5):1223– 1229
- ↵Tommiska V, Lehtonen L, Renlund M, et al. No improvement in outcome of nationwide extremely low birth weight infant population between 1996–1997 and 1999–2000. Pediatrics.2007;119 (1):29– 36
- ↵Hamilton BE, Minino AM, Martin JA, et al. Annual summary of vital statistics: 2005. Pediatrics.2007;119 (2):345– 360
- ↵Kawamoto Y, Fujimura M, Ogawa Y. Is the incidence of chronic lung disease decreasing? The National CLD survey in Japan: changes over 10 years [in Japanese]. J Jpn Soc Premature Newborn Med.2002;14 (3):352 (abstract)
- ↵Vohr BR, Wright LL, Dusick AM, et al. Neurodevelopmental and functional outcome of extremely low birth weight infants in the National Institute of Child Health and Human Development Neonatal Research Network, 1993–1994. Pediatrics.2000;105 (6):1216– 1226
- Copyright © 2009 by the American Academy of Pediatrics