Population-Based Analyses of Mortality in Trisomy 13 and Trisomy 18
Sonja A. Rasmussen, MD, MS*; Lee-Yang C. Wong, MS‡; Quanhe Yang, PhD*; Kristin M. May, PhD§; and J. M. Friedman, MD, PhD储
ABSTRACT. Objective. Although trisomy 13 and tri-somy 18 are generally considered to be lethal, long-term survival of patients has been reported. We sought to evaluate mortality in people with trisomy 13 or 18 using 2 population-based strategies.
Methods. In the first analysis, infants who had tri-somy 13 or 18 and were born during 1968 –1999 were identified using the Metropolitan Atlanta Congenital Defects Program, a population-based birth defects sur-veillance system. Dates of death were documented using hospital records, Georgia vital records, and the National Death Index. In the second analysis, we used the Multi-ple-Cause Mortality Files compiled from US death cer-tificates from 1979 through 1997. Using these 2 analyses, we examined median survival time or median age at death, survival beyond 1 year of age, and factors associ-ated with longer survival.
Results. Using Metropolitan Atlanta Congenital De-fects Program, we identified 70 liveborn infants with trisomy 13 and 114 liveborn infants with trisomy 18. Median survival time was 7 days (95% confidence inter-val [CI]: 3–15) for people with trisomy 13 and 14.5 days (95% CI: 8 –28) for people with trisomy 18. For each con-dition, 91% of infants died within the first year. Neither race nor gender affected survival for trisomy 13, but for trisomy 18, girls and infants of races other than white seemed to survive longer. The presence of a heart defect did not seem to affect survival for either condition. Using MCMF, we identified 5515 people with trisomy 13 and 8750 people with trisomy 18 listed on their death certif-icates. Median ages at death for people with trisomy 13 and trisomy 18 both were 10 days; 5.6% of people with trisomy 13 and 5.6% of people with trisomy 18 died at age 1 year or greater. Race and gender seemed to affect sur-vival in both conditions, with girls and blacks showing higher median ages at death.
Conclusions. Although survival is greatly affected by trisomy 13 and trisomy 18, 5% to 10% of people with these conditions survive beyond the first year of life. These population-based data are useful to clinicians who care for patients with these trisomies or counsel families with infants or fetuses who have a diagnosis of trisomy 13 or 18.Pediatrics2003;111:777–784;trisomy 13, trisomy 18, survival, mortality.
ABBREVIATIONS. MACDP, Metropolitan Atlanta Congenital De-fects Program; MCMF, Multiple-Cause Mortality Files;ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Mod-ification; NDI, National Death Index; CI, confidence interval;
ICD-9, International Classification of Diseases, Ninth Revision.
T
risomy 18 and trisomy 13 are, respectively, the second and third most commonly diagnosed autosomal trisomies in liveborn infants. Al-though these conditions are associated with a high degree of infant mortality,1longer survival has been reported.2–16Studies have reported median survival times varying from 2.5 days17 to between 1 and 4 months18 in trisomy 13, and from 2.5 days19 to 70 days20,21 in trisomy 18. Some of these studies ascer-tained patients through personal communication or literature reports and this may have introduced a bias toward longer survival. A study based on ques-tionnaires that were sent to members of a support group for families of people with trisomy 13 and 18 reported even longer survival: 38% of people with trisomy 13 and 42% with trisomy 18 were still alive at age 1 year. However, as acknowledged by the study authors, that study was likely to be strongly biased toward longer survival, because families with infants who died shortly after birth would be less likely to be members of the support group. Although some of the reported variability in survival may be the result of differences in diagnosis or patient treatment,19 differences in case ascertainment methods also seem to play a significant role. Thus, studies that use pop-ulation-based methods of ascertainment can provide much-needed accurate data on mortality of people with trisomy 13 or 18.Information on factors associated with long-term survival are limited, but several studies have sug-gested that girls with trisomy 13 or 18 live longer than boys.20 –25 Factors other than gender have not been well studied. For example, despite the recent finding that survival of people with Down syndrome is greatly affected by race,26 the effects of race on trisomy 13 or 18 survival have not been carefully studied. Some authors have suggested that the ab-sence of heart malformations may be associated with a longer lifespan,2,27 but others have noted that the types of heart defects most commonly associated with these conditions would not be expected to be lethal in infancy, even if not surgically treated.28 Some authors18have suggested that certain cytoge-netic forms (translocations and mosaics) may be as-sociated with longer survival than free trisomies, but others have not noted this difference for
transloca-From the *National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention, Atlanta, Georgia; ‡Health Investigations Branch, Division of Health Studies, Agency for Toxic Sub-stances and Disease Registry, Atlanta, Georgia; §Department of Pediatrics, Division of Medical Genetics, Emory University School of Medicine, At-lanta, Georgia; and储Department of Medical Genetics, University of British Columbia, Vancouver, British Columbia, Canada.
Received for publication Apr 30, 2002; accepted Sep 4, 2002.
Reprint requests to (S.A.R.) 4770 Buford Hwy NE, CDC, MS F-45, Atlanta, GA 30341. E-mail: [email protected]
tions.29Some have suggested that long-term survival of infants with trisomy 13 or 18 is associated with more aggressive management and have proposed that the reason that recent studies show decreased survival is because infants who have trisomy 13 or 18 and are born in recent years receive less aggressive care because caregivers expect them to die very young.17,30 However, other authors have noted that long-term survivors have not received particularly aggressive or extraordinary care.31Accurate descrip-tion of mortality and understanding of factors asso-ciated with longer survival are important to clini-cians who care for infants with these trisomies or counsel families with fetuses or infants who have a diagnosis of trisomy 18 or 13.
We sought to evaluate mortality in people with trisomy 13 or trisomy 18 using 2 population-based strategies. In the first, liveborn infants with trisomy 13 or 18 were identified through a birth defects sur-veillance system, the Metropolitan Atlanta Congeni-tal Defects Program (MACDP). We calculated me-dian age of survival and assessed the impact of factors such as race, gender, presence of heart defect, and period of birth. We compared these data with those obtained from our second data source, Multi-ple-Cause Mortality Files (MCMF), compiled from US death certificates.
METHODS MACDP
We ascertained infants who were born with trisomy 13 or trisomy 18 during 1968 –1999 using MACDP, a population-based birth defects surveillance program. Since 1968, MACDP has ac-tively monitored birth defects among fetal deaths and live infants who were born to women who resided within a 5-county metro-politan Atlanta area. The program routinely collects clinical infor-mation and demographic characteristics on infants with major birth defects. Trained abstractors collect information from birth and pediatric hospitals, cytogenetic laboratories, and a referral center that provides services to children with congenital heart defects in the area. MACDP data include dates of death for some people, primarily those who died while at a hospital in the 5-county area. Additional details about MACDP have been pub-lished elsewhere.32
To identify infants with trisomy 13 or 18, we used codes based on theInternational Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), 758.100 –758.190 for trisomy 13 and 758.200 –758.290 for trisomy 18. We reviewed available cytogenetic data for all liveborn infants coded as having trisomy 13 or trisomy 18 and included only infants whose diagnosis was cytogenetically confirmed. Infants with free trisomy or a Robertsonian transloca-tion were included, but infants with mosaicism and those with partial duplication of chromosomes 13 or 18 were excluded. We classified infants as having a heart defect when MACDP listed them as having a code for a heart defect (ICD-9-CM-based codes 745.000 –747.900). For our analyses, infants with minor cardiac defects (eg, patent foramen ovale, patent ductus arteriosus, tricus-pid insufficiency) or unconfirmed cardiac defects were not classi-fied as having a heart defect.
When possible, we identified deaths using data from MACDP and from Georgia vital records. For people whose date of death was unknown, we supplemented these data by linking MACDP cases with deaths listed in the National Death Index (NDI), a centralized index of death record information for 1979 through 1998, compiled by the National Center for Health Statistics. Details on the NDI matching process are described elsewhere.33,34 For people who had trisomy 13 or 18 and had no death record in MACDP, Georgia vital records, or NDI, survival was censored at the end of the follow-up period on December 31, 1999. People for whom no date of death was found were searched for among
admission records to local pediatric hospitals to determine the latest age that the person was documented as being alive.
For people with trisomy 13 or trisomy 18, we estimated the survival probability, including the median survival time, from birth by the Kaplan-Meier product-limit method35for the entire study period. We also estimated median survival time and Kaplan-Meier survival probability by period of birth (1968 –1979, 1980 –1989, and 1990 –1999) and for possible prognostic factors, including gender, race, and presence of a heart defect. We used Greenwood’s method to calculate the 95% confidence intervals (CIs) for the estimate of the survival probability and the sign test method to calculate CIs for median survival time.36We used the log-rank test (SAS Institute, Cary, NC) to examine variation in survival by period of birth and possible prognostic factors.
MCMF
We used MCMF data compiled by the National Center for Health Statistics for 1979 through 1997. MCMF data include de-mographic information andInternational Classification of Diseases, Ninth Revision (ICD-9)codes for the underlying cause of death and up to 20 conditions listed on the death certificates of⬎40 million people who died in the United States during this period.37Death certificates record death events only among live births; stillborn infants are not included.38MCMF data exist in entity axis format and record axis format. The entity axis format provides a separate code for each disease listed on the death certificate whether it is an underlying cause or a contributory condition. The record axis format uses linkage rules to combine some listings of conditions, to determine the underlying cause of death and other contributory conditions and their positions as listed on the death certificates.37 We selected all records that contained code 758.1 (trisomy 13) or 758.2 (trisomy 18) anywhere in the record axis portion. From these records, we excluded records that contained code 779.6 (pregnan-cy termination).
We calculated median age at death over time for people with trisomy 13 or with trisomy 18 by race (“white,” “black,” or “oth-er”) and gender. An infant was classified as having a heart defect when any heart defect code (745.0 –747.9) was listed on the death certificate. We used linear regression to test the trend of median age at death by year and used the nonparametric median scores method to test differences of median age at death by racial groups.39
RESULTS Trisomy 13
MACDP
During the period 1968 –1999, 83 liveborn infants were coded as having trisomy 13. After review of the cytogenetic data, we excluded 13 of these infants— 6 because their cytogenetic results were not available, 2 because they had mosaicism, and 5 because they had other abnormalities involving chromosome 13. After these infants were excluded, we had data on 70 liveborn infants, including 9 with Robertsonian translocations.
We identified the deaths of 69 of the 70 people with trisomy 13. MACDP records captured 56 deaths, linkage with Georgia vital records identified 12 additional deaths, and NDI linkage identified 1 additional death. Of the 70 people with trisomy 13, 64 (91%) died during the first year of life. Five died at ages⬎1 year (383, 791, 818, 1000, and 1858 days). The child for whom no date of death was available was last documented as being alive at 1117 days of age, based on local pediatric hospital records.
CI: 3–15) and did not vary significantly by any of the selected clinical or demographic factors (Table 1). Probabilities of survival to 1 month and 1 year did not vary significantly by gender, race, or presence of heart defect (Table 2). However, the 1-month sur-vival probabilities varied significantly by birth pe-riod, from 40% (1968 –1979) to 47% (1980 –1989) to 17% (1990 –1999). One-year survival probabilities did not vary significantly between time periods.
MCMF
MCMF contain records for 40 591 313 deaths dur-ing 1979 –1997, 5553 of which listed trisomy 13 on the death certificate. Of these, we excluded 6 records that also listed the code for pregnancy termination, 30 that listed codes for both trisomy 13 and 18, and 2 that were missing values for age at death. The final analytical death cohort for trisomy 13 was 5515, in-dicating that 1 in 7362 deaths was associated with trisomy 13.
The racial distribution among trisomy 13-associ-ated deaths was 78.7% white, 18.4% black, and 3.5% others, whereas the racial distribution of all births for 1980 –1997 was 79.9% white, 15.9% black, and 4.2%
others.40The overall median age at death for the 5515 people with trisomy 13 was 10 days (5th percentile, 1 day; 25th percentile, 1 day; 75th percentile, 30 days; 95th percentile, 1365 days). Of trisomy 13-associated deaths, 1677 (30.4%) occurred at 1 month or greater and 309 (5.61%) occurred at age 1 year or greater. The median age at death decreased from 10 days in 1979 to 6 days in 1997 (⫾standard error⫽ ⫺0.67⫾0.23;
P⫽.056). The median age at death was significantly higher for girls than for boys (10 days vs 5 days;P⬍
.0001; Fig 2), and blacks had a higher median age at death (17 days) than whites (6 days) or people of races other than white or black (5 days;P⬍ .0001). The median age at death was not significantly dif-ferent between whites and people of other races (ex-cluding blacks). Median age at death was higher for people with a heart defect noted on the death certif-icate (10 days) than for those without a heart defect noted (6 days;P ⬍.0001).
Trisomy 18
MACDP
For 1968 through 1999, MACDP coded 124 live-born infants as having trisomy 18. After reviewing the cytogenetic data, we excluded 10 of these in-fants—3 because cytogenetic results were not avail-able, 3 because they had mosaicism, and 4 because they had other abnormalities involving chromosome 18. After these infants were excluded, 114 liveborn infants with trisomy 18 remained for study, includ-ing 4 infants who also had sex chromosome aneu-ploidy (3 with 48, XXY, ⫹18 and 1 with 48, XXX, ⫹18) and 2 with free trisomy 18 and an additional apparently balanced chromosomal rearrangement.
We identified deaths in 106 of the 114 people with trisomy 18. MACDP records captured 83 deaths, linkage with Georgia vital records identified 22 ad-ditional deaths, and NDI linkage identified 1 addi-tional death. Of the 114 people with trisomy 18, 104 (91%) died during the first year of life. Two people with trisomy 18 died at ages⬎1 year (414 days and 1081 days). Of the 8 children with trisomy 18 for whom no date of death was available, 5 were last documented by pediatric hospital records as being alive at ages 410, 810, 1332, 1848, and 3203 days. For the other 3 (1 born in 1997 and 2 in 1999), no addi-tional information was available.
For liveborn infants with trisomy 18, the estimated probability of survival to 1 month of age was 38.6% (95% CI: 29.7– 47.5) and to 1 year of age was 8.4% (95% CI: 3.2–13.6; Fig 3). Median survival time was 14.5 days (95% CI: 8 –28). Analysis of median sur-vival time showed significantly longer sursur-vival for girls (27 days) compared with boys (3 days). The median survival time was longest for Hispanic peo-ple, compared with black or white people. The me-dian survival time of infants with trisomy 18 in-creased from 10 days (1968 –1979) to 14 days (1980 – 1989) and then to 19 days (1990 –1999), but this increase was not statistically significant. The pres-ence of a congenital heart defect was not associated with significantly shorter lifespan (Table 3). Analysis of 1-month survival probabilities for infants with
Fig 1. Kaplan-Meier survival curve (solid line) and 95% CIs (dashed lines) for children with trisomy 13, metropolitan Atlanta, 1968 –1999 (truncated at 1 year).
TABLE 1. Median Survival Time of People With Trisomy 13 by Selected Demographic and Clinical Characteristics, Atlanta, 1968 –1999
Characteristics No. of Births (n⫽70)
No. of Deaths (% Died)
Median Survival Time (Day; 95% CI)
Race*
White 35 35 (100) 7 (2–26)
Black 33 32 (97) 10 (2–19)
Hispanic 1 1 (100)
Gender†
Male 36 36 (100) 10 (5–19)
Female 33 32 (97) 7 (2–23)
Heart defect
Present 32 32 (100) 10 (5–26)
Absent 38 37 (97) 6 (2–15)
Period of birth
1968–1979 15 15 (100) 10 (2–42)
1980–1989 21 21 (100) 26 (5–59)
1990–1999 34 33 (97) 2.5 (2–11)
trisomy 18 by selected demographic and clinical characteristics revealed significant differences asso-ciated with race and gender (Table 4). The 1-month survival probability was significantly higher for blacks than for whites (P ⫽ .0213) when the three
Hispanic people were excluded. Girls were signifi-cantly more likely to survive 1 month than boys, although no significant gender difference was noted for 1-year survival probabilities.
Fig 2. Median age at death among 5514 liveborn indi-viduals with trisomy 13 by gender, MCMF data, 1979 – 1997.
Fig 3. Kaplan-Meier survival curve (solid line) and 95% CIs (dashed lines) for children with trisomy 18, metropolitan Atlanta, 1968 –1999 (truncated at 1 year).
TABLE 2. 1-Month and 1-Year Survival Probabilities of People With Trisomy 13 by Selected Demographic and Clinical Characteristics, Atlanta, 1968 –1999
Characteristics 1-Month Survival Probability
(95% CI)
PValue for 1-Month Survival
1-Year Survival Probability
(95% CI)
PValue for 1-Year Survival
Race*
White 0.31 (0.16–0.47) .5322 0.085 (0.00– 0.18)
.9780
Black 0.27 (0.12–0.42) 0.09 (0.00–0.42)
Gender†
Male 0.31 (0.16–0.46) .7513 0.08 (0.00–0.17) .6842
Female 0.30 (0.17–0.46) 0.09 (0.00–0.19)
Heart defect
Present 0.31 (0.15–0.47) .5541 0.16 (0.04–0.27) .4785
Absent 0.32 (0.19–0.46) 0.0 (0.00–0.00)
Period of birth
1968–1979 0.40 (0.15–0.64) 0.06 (0.00–0.19)
1980–1989 0.47 (0.26–0.68) .0310 0.14 (0.00–0.29) .0849 1990–1999 0.17 (0.05–0.30) 0.06 (0.00–0.14)
* One person of unknown race. † One person of unknown gender.
TABLE 3. Median Survival Time of People With Trisomy 18 by Selected Demographic and Clinical Characteristics, Atlanta, 1968 –1999
Characteristics No. of Births (n⫽114)
No. of Deaths (% Died)
Median Survival Time (Day; 95% CI)
Race
White 59 56 (95) 4 (2–18)
Black 51 48 (94) 24 (13–43)
Hispanic 3 1 (33) 275 (62–414)
Asian 1 0 (0)
Gender
Male 40 37 (93) 3 (1–8)
Female 74 69 (93) 27 (14–39)
Heart defect
Present 67 61 (91) 14 (6–28)
Absent 47 45 (96) 20 (4–34)
Period of birth
1968–1979 20 20 (100) 10 (3–61)
1980–1989 46 45 (98) 14 (7–34)
MCMF
MCMF for 1979 –1997 contained 8797 death certif-icate records that listed trisomy 18. Of those, we excluded 15 that also listed the code for pregnancy termination, 30 that listed codes for both trisomy 13 and 18, and 2 that were missing values for age at death. The final analytical death cohort for trisomy 18 was 8750, indicating that 1 in 4639 deaths was associated with trisomy 18.
The racial distribution among people with trisomy 18 was 78.9% white, 16.5% black, and 4.6% other races. Among people with trisomy 18, the overall median age at death was 10 days (5th percentile, 1 day; 25th percentile, 2 days; 75th percentile, 60 days; 95th percentile, 1365 days). Among trisomy 18-asso-ciated deaths, 3316 (37.9%) occurred at 1 month or greater and 487 (5.57%) occurred at age 1 year or greater. The median age at death decreased from 17 days in 1979 to 10 days in 1997 (⫾standard error⫽ ⫺0.62 ⫾ 0.17; P ⫽ .0003). The median age at death was significantly higher among girls than among boys (17 days vs 5 days, respectively;P⬍.0001), and
blacks had a higher median age at death (17 days) than whites (10 days) or people of other races (10 days;P⬍.0001; Fig 4). The median age at death for people with trisomy 18 with a heart defect noted on the death certificate was higher (17 days) than for those without a heart defect noted (10 days; P ⬍
.0001).
DISCUSSION
We used 2 different population-based strategies to study mortality in trisomy 13 and trisomy 18. The 2 data sets used in this study have important strengths. Both are population-based, which decreases the se-lection bias that has been observed in some studies. Because MACDP has been in operation since 1968, these data made it possible for us to evaluate changes in survival among people ascertained with identical criteria over a period of 32 years. In addition, MACDP’s multiple ascertainment sources, including data from cytogenetic laboratories,41 allowed us to exclude people whose condition was not cytogenet-ically confirmed. In most cases, exclusions were
Fig 4. Median age at death among 8750 liveborn indi-viduals with trisomy 18 by gender, MCMF data, 1979 – 1997.
TABLE 4. 1-Month and 1-Year Survival Probabilities of People With Trisomy 18 by Selected Demographic and Clinical Characteristics, Atlanta, 1968 –1999
Characteristics 1-Month Survival Probability
(95% CI)
PValue of 1-Month Survival
1-Year Survival Probability
(95% CI)
PValue of 1-Year Survival
Race
White 0.28 (0.17–0.40) 0.04 (0.00–0.10)
Black 0.45 (0.31–0.59) .0211 0.08 (0.01–0.15) .0113
Hispanic 1.00 (1.0–1.0) 0.66 (0.13–1.00)
Asian 1.00 (1.0–1.0) 1.00 (1.0–1.0)
Gender
Male 0.28 (0.17–0.41) .0035 0.09 (0.00–0.19) .0695
Female 0.45 (0.33–0.56) 0.08 (0.02–0.14)
Heart defect
Present 0.37 (0.26–0.49) .9089 0.12 (0.04–0.20) .5960
Absent 0.40 (0.26–0.54) 0.04 (0.00–0.10)
Period of birth
1968–1979 0.35 (0.14–0.55) 0.00 (0.00–0.00)
1980–1989 0.37 (0.23–0.51) .8960 0.04 (0.00–0.10) .3562
made because another cytogenetic abnormality or mosaicism was diagnosed, not because of lack of cytogenetic data. Use of NDI to provide information on people whose date of death was not available from hospital records or Georgia death certificates is another strength of our study because previous stud-ies have shown that information obtained from NDI data is accurate and that this source identifies a high proportion of deaths.42– 44MCMF data allowed us to assess much larger numbers of people than previous studies.
Both data sets also impose limitations. The biggest limitation with MACDP data is that people with trisomy 13 or 18 for whom no date of death was available in the hospital records, Georgia vital records, or NDI were assumed to be alive, but this assumption may be incorrect. Linkage to Georgia vital records depended on name; thus, name changes complicated this linkage. However, nearly all people for whom no date of death was found were born during the time period for which NDI data are avail-able, and NDI uses several factors for linkage. In addition, local pediatric hospital records docu-mented 6 of the 9 people with an unknown date of death as being alive after the age of 1 year. Finally, median age of survival is unlikely to be severely affected by these few people. Another limitation of MACDP data is that, because it considered only in-fants whose conditions were cytogenetically con-firmed, it missed infants who had trisomy 13 or 18 and died before blood or other samples were ob-tained for cytogenetic analysis. Thus, our use of these data may skew our findings toward longer survival. The major limitation of MCMF data is that they are based on death certificates, which previous studies have demonstrated to be incomplete and, in some instances, inaccurate.45,46 In this data set, we were unable to confirm diagnoses of trisomy 13 or 18 by cytogenetic analysis, so other conditions may have been incorrectly diagnosed and recorded as trisomy 13 or 18. This problem is highlighted by the finding that 30 records listed both trisomy 13 and trisomy 18 on the death certificate (these records were excluded from our analysis). Another limitation is that we could not exclude people with mosaicism, who may be expected to have a longer survival. In addition, people who had trisomy 13 or 18 and were not diagnosed before their death certificates were com-pleted would not be included in MCMF data. Data on the presence of a heart defect are limited by our use of death certificates as the source of information. We were unable to assess the accuracy of ascer-tainment of trisomy 13 or 18 in the MCMF data. In a study from Hawaii, the prevalence of trisomy 13 among liveborns was 1 in 12 083 and for trisomy 18 was 1 in 655927; MACDP data for 1968 –1999 show the frequency of cytogenetically confirmed cases to be 1 in 14 700 for trisomy 13 and 1 in 9026 for trisomy 18. These figures for frequency of trisomy 13 and trisomy 18 are lower than those reported on death certificates. One reason for this may be that the Ha-waii study included pregnancy terminations; during the study period (1986 –1997), nearly 40% of trisomy 13 and almost 50% of trisomy 18 cases were
preg-nancy terminations. Our study using MCMF data included earlier years, a time when prenatal diagno-sis and pregnancy termination would have been less frequent. The birth prevalence of trisomy 13 or tri-somy 18 in Hawaii would have been 1 in 6205 and 1 in 2503, respectively, if elective terminations were included. Another reason for the higher frequency of trisomy 13 and trisomy 18 in MCMF data may be that results of chromosome analyses were not available when the death certificates were completed, so some infants may have been incorrectly classified as hav-ing 1 of these conditions.
Another limitation of our analyses is that we have no information on the types of medical interventions provided to people with trisomy 13 or 18 in either data set, although it is likely that a wide range of medical interventions were provided. Because of this, we are unable to address whether people who have trisomy 13 or 18 and survive longer receive more aggressive care.
Despite these limitations, the 2 data sets showed similar results for median survival time (MACDP) and median age at death (MCMF). The median sur-vival time for infants with trisomy 13 was 7 days using MACDP data, and the median age at death was 10 days using MCMF data. The median survival time for infants with trisomy 18 was 14.5 days using MACDP data, and the median age at death was 10 days using MCMF data. These findings were similar to those obtained from other unselected studies of mortality, which reported median survival of 2.517 and 428days for people with trisomy 13 and median survival ranging from 2.5 to 6 days for people with trisomy 18.17,19,21,24,25
Using data from MACDP, we found that 91% of infants with trisomy 13 or with trisomy 18 died within the first year of life. This is also similar to previous unselected studies. The range in previous studies for trisomy 13 was 89.5% to 100%,17,28,47and for trisomy 18 was 74.3% to 100%.17,19,24,47 One of these studies showed lower rates of death in the first year of life47; however, this study may have under-estimated infant mortality because deaths that oc-curred out of state may have been underascertained. Like our study, previous studies have suggested that girls with trisomy 13 or 18 survive longer than boys.20 –25 Using MCMF data, we observed signifi-cantly higher median age at death for girls than for boys with either trisomy 13 or trisomy 18. Similarly, using MACDP data, we found that girls with trisomy 18 had longer median survival and significantly bet-ter 1-month survival probability than boys. Longer survival in girls was not seen in trisomy 13 using MACDP data.
was previously suggested when 2 black children with trisomy 13 were documented to have survived into the second decade of life.30
MACDP data suggest that the presence of a heart defect does not negatively affect survival for either condition. MCMF data indicate that people who had trisomy 13 or 18 and also had a heart defect listed on their death certificates survived longer than those without a heart defect listed. These data could be helpful in the discussion regarding whether cardiac surgery is indicated in people with trisomy 13 or 1831; however, these data are likely to be affected by the fact that many infants with either trisomy 13 or trisomy 18 die very early in life, before a heart defect is diagnosed. This conjecture is supported by the fact that the proportion of people who had trisomy 13 or 18 and heart defects in our study is smaller than previously reported.48
The results of our examination of mortality by period of birth are difficult to interpret. Using MCMF data, we found that the median age at death signif-icantly decreased through the years studied for both trisomy 13 and trisomy 18. MACDP data showed a significant trend over the 3 time periods of birth for shorter 1-month survival probability in people with trisomy 13 but not trisomy 18. No statistically signif-icant differences by time period were seen for 1-year survival probability or median survival time for peo-ple with trisomy 13 or 18 or for 1-month survival probability for people with trisomy 18. Recent stud-ies have reported shorter survival than earlier ones, leading some authors to suggest that patients are dying younger, possibly because infants who have trisomy 13 or 18 and are born in recent years receive less aggressive care because of the expectation that they will die early. Differences in case ascertainment methods are also likely to affect reported survival.
These data demonstrate the pattern of mortality associated with trisomy 13 and with trisomy 18. De-spite the poor survival of people with these condi-tions, it is important to recognize that 5% to 10% of children with trisomy 13 or trisomy 18 do survive the first year. This information is important to providers who care for patients with these conditions and their families. Given our focus on liveborns, our data should be used cautiously for prenatal counseling because a significant proportion of these pregnancies are expected to be lost before birth.49
ACKNOWLEDGMENTS
We thank MACDP staff members Debra Adams, Fran Baxter, Jo Anne Croghan, Joann Donaldson, Joan Garcia, Debbie Nurmi, Mary Kathryn Peecher, Charlie Mae Peters, Wendy Sklenka, Car-olyn Sullivan, Karen Thornton, and Tineka Yowe. Their constant data collection efforts provide the foundation on which MACDP is built. We also acknowledge Dr Leslie O’Leary for assistance in obtaining additional information on MACDP study patients.
REFERENCES
1. Smith DW. Autosomal abnormalities. Am J Obstet Gynecol.1964;90: 1055–1077
2. Zoll B, Wolf J, Lensing-Hebben D, Pruggmayer M, Thorpe B. Trisomy 13 (Patau syndrome) with an 11-year survival.Clin Genet.1993;43:46 –50 3. Smith A, Field B, Learoyd BM. Trisomy 18 at age 21 years.Am J Med
Genet.1989;34:338 –339
4. Smith A, Silink M, Ruxton T. Trisomy 18 in an 11 year old girl.J Ment Defic Res.1978;22:277–286
5. Surana RB, Bain HW, Conen PE. 18-Trisomy in a 15-year-old girl.Am J Dis Child.1972;123:75–77
6. Karayalcin G, Shanske A, Honigman R. Wilms’ tumor in a 13-year old girl with trisomy 18.Am J Dis Child.1981;135:665– 666
7. Mankinen CB, Sears JW. Trisomy 13 in a female over 5 years of age.
J Med Genet.1976;13:157–161
8. Cowen JM, Walker S, Harris F. Trisomy 13 and extended survival.J Med Genet.1979;16:155–157
9. Mehta L, Shannon RS, Duckett DP, Young ID. Trisomy 18 in a 13 year old girl.J Med Genet.1986;23:256 –257
10. Singh KS. Trisomy 13 (Patau’s syndrome): a rare case of survival into adulthood.J Ment Defic Res.1990;34:91–93
11. Ray S, Ries MD, Bowen JR. Arthrokatadysis in trisomy 18.J Pediatr Orthop.1986;6:100 –102
12. Geiser CF, Schindler AM. Long survival in a male with 18-trisomy syndrome and Wilms’ tumor.Pediatrics.1969;44:111–116
13. Redheendran R, Neu RL, Bannerman RM. Long survival in trisomy-13-syndrome: 21 cases including prolonged survival in two patients 11 and 19 years old.Am J Med Genet.1981;8:167–172
14. Van Dyke DC, Allen M. Clinical management considerations in long-term survivors with trisomy 18.Pediatrics.1990;85:753–759
15. Weber WW, Mamunes P, Day R, Miller P. Trisomy 17–18 (E): studies in long-term survival with report of two autopsied cases.Pediatrics.1964; 34:533–541
16. Hook EB, Lehrke R, Roesner A, Yunis JJ. Trisomy-18 in a 15-year-old female.Lancet.1965;2:910 –911
17. Goldstein H, Nielsen KG. Rates and survival of individuals with tri-somy 13 and 18. Data from a 10-year period in Denmark.Clin Genet.
1988;34:366 –372
18. Magenis RE, Hecht F, Milham S Jr. Trisomy 13 (D1) syndrome: studies on parental age, sex ratio, and survival.J Pediatr.1968;73:222–228 19. Young ID, Cook JP, Mehta L. Changing demography of trisomy 18.Arch
Dis Child.1986;61:1035–1036
20. Weber WW. Survival and the sex ratio in trisomy 17–18.Am J Hum Genet.1967;19:369 –377
21. Root S, Carey JC. Survival in trisomy 18.Am J Med Genet.1994;49: 170 –174
22. Baty BJ, Blackburn BL, Carey JC. Natural history of trisomy 18 and trisomy 13: I. Growth, physical assessment, medical histories, survival, and recurrence risk.Am J Med Genet.1994;49:175–188
23. Conen PE, Erkman B. Frequency and occurrence of chromosomal syn-dromes. II. E-Trisomy.Am J Hum Genet.1966;18:387–398
24. Embleton ND, Wyllie JP, Wright MJ, Burn J, Hunter S. Natural history of trisomy 18.Arch Dis Child Fetal Neonatal Ed.1996;75:F38 –F41 25. Carter PE, Pearn JH, Bell J, Martin N, Anderson NG. Survival in trisomy
18. Life tables for use in genetic counselling and clinical paediatrics.Clin Genet.1985;27:59 – 61
26. Racial disparities in median age at death of persons with Down syn-drome—United States, 1968 –1997.MMWR Morb Mortal Wkly Rep.2001; 50:463– 465
27. Taylor AI. Autosomal trisomy syndromes: a detailed study of 27 cases of Edwards’ syndrome and 27 cases of Patau’s syndrome.J Med Genet.
1968;5:227–252
28. Wyllie JP, Wright MJ, Burn J, Hunter S. Natural history of trisomy 13.
Arch Dis Child.1994;71:343–345
29. Hodes ME, Cole J, Palmer CG, Reed T. Clinical experience with tri-somies 18 and 13.J Med Genet.1978;15:48 – 60
30. Hecht F. Who will survive with trisomy 13 or 18? A call for cases 10 years old or above.Am J Med Genet.1981;10:417– 418
31. Carey JC. Trisomy 18 and trisomy 13 syndromes. In: Cassidy S, Allan-son J, eds.Management of Genetic Syndromes. New York, NY: Wiley-Liss; 2001:417– 436
32. Edmonds LD, Layde PM, James LM, Flynt JW, Erickson JD, Oakley GP Jr. Congenital malformations surveillance: two American systems.Int J Epidemiol.1981;10:247–252
33. Wentworth DN, Neaton JD, Rasmussen WL. An evaluation of the Social Security Administration master beneficiary record file and the National Death Index in the ascertainment of vital status.Am J Public Health.
1983;73:1270 –1274
34. National Center for Health Statistics.National Death Index User Manual. Hyattsville, MD: National Center for Health Statistics; 2000
35. Kaplan EL, Meier P. Nonparametric estimation from incomplete obser-vations.J Am Stat Assoc.1958;53:457– 481
36. Brookmeyer R, Crowley J. A confidence interval for the median survival time.Biometrics.1982;38:29 – 41
cause-of-death data.Am J Epidemiol.1986;124:161–179
38. National Center for Health Statistics. Public Use Data Tape Documentation: Multiple Cause of Death for ICD-9 Data. 1982 edition. Washington, DC: US Government Printing Office; 1986
39. Conover WJ.Practical Nonparametric Statistics.3rd ed. New York, NY: Wiley; 1998
40. Martin JA, Hamilton BE, Ventura SJ, Menacker F, Park MM. Births: final data for 2000.Natl Vital Stat Rep.2002;50:1–101
41. Honein MA, Paulozzi LJ. Birth defects surveillance: assessing the “gold standard.”Am J Public Health.1999;89:1238 –1240
42. Rich-Edwards JW, Corsano KA, Stampfer MJ. Test of the National Death Index and Equifax Nationwide Death Search.Am J Epidemiol.
1994;140:1016 –1019
43. Boyle CA, Decoufle P. National sources of vital status information: extent of coverage and possible selectivity in reporting.Am J Epidemiol.
1990;131:160 –168
44. Lash TL, Silliman RA. A comparison of the National Death Index and
Social Security Administration databases to ascertain vital status. Epi-demiology.2001;12:259 –261
45. Lloyd-Jones DM, Martin DO, Larson MG, Levy D. Accuracy of death certificates for coding coronary heart disease as the cause of death.Ann Intern Med.1998;129:1020 –1026
46. Sirken MG, Rosenberg HM, Chevarley FM, Curtin LR. The quality of cause-of-death statistics.Am J Public Health.1987;77:137–139 47. Forrester MB, Merz RD. Trisomies 13 and 18: prenatal diagnosis and
epidemiologic studies in Hawaii, 1986 –1997. Genet Test. 1999;3: 335–340
48. Musewe NN, Alexander DJ, Teshima I, Smallhorn JF, Freedom RM. Echocardiographic evaluation of the spectrum of cardiac anomalies associated with trisomy 13 and trisomy 18.J Am Coll Cardiol.1990;15: 673– 677
49. Hook EB, Cross PK, Schreinemachers DM. Chromosomal abnormality rates at amniocentesis and in live-born infants. JAMA. 1983;249: 2034 –2038
MEDICAL DIVERSIONS
“Every country in the developing world is increasing its expenditure on health care in. . . an unattainable battle against death, pain, and sickness. More and more of life’s processes and difficulties— birth, death, sexuality, aging, unhappiness, tiredness, loneliness, (and) perceived imperfections in our bodies are being medi-calized. Medicine cannot solve these problems. It can sometimes help, but often at a substantial cost. People become patients. Stigma proliferates. Large sums are spent. The treatments may be poisonous and disfiguring. Worst of all, people are diverted from what may be much better ways to adjust to their problems.”
Smith R. Spend (slightly) less on health and more on arts.BMJ.2002;325:1432–1433
DOI: 10.1542/peds.111.4.777
2003;111;777
Pediatrics
Friedman
Sonja A. Rasmussen, Lee-Yang C. Wong, Quanhe Yang, Kristin M. May and J. M.
Population-Based Analyses of Mortality in Trisomy 13 and Trisomy 18
Services
Updated Information &
http://pediatrics.aappublications.org/content/111/4/777
including high resolution figures, can be found at:
References
http://pediatrics.aappublications.org/content/111/4/777#BIBL
This article cites 45 articles, 11 of which you can access for free at:
Subspecialty Collections
http://www.aappublications.org/cgi/collection/genetics_sub
Genetics
following collection(s):
This article, along with others on similar topics, appears in the
Permissions & Licensing
http://www.aappublications.org/site/misc/Permissions.xhtml
in its entirety can be found online at:
Information about reproducing this article in parts (figures, tables) or
Reprints
http://www.aappublications.org/site/misc/reprints.xhtml
DOI: 10.1542/peds.111.4.777
2003;111;777
Pediatrics
Friedman
Sonja A. Rasmussen, Lee-Yang C. Wong, Quanhe Yang, Kristin M. May and J. M.
Population-Based Analyses of Mortality in Trisomy 13 and Trisomy 18
http://pediatrics.aappublications.org/content/111/4/777
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.
