Research

JAMA Pediatrics | Original Investigation

Developmental Variability in Autism Across 17 000 Autistic Individuals

and 4000 Siblings Without an Autism Diagnosis
Comparisons by Cohort, Intellectual Disability,
Genetic Etiology, and Age at Diagnosis
Susan S. Kuo, PhD; Celia van der Merwe, PhD; Jack M. Fu, PhD; Caitlin E. Carey, PhD;
Michael E. Talkowski, PhD; Somer L. Bishop, PhD; Elise B. Robinson, DSc

Supplemental content
IMPORTANCE Presence of developmental delays in autism is well established, yet few studies
have characterized variability in developmental milestone attainment in this population.

OBJECTIVE To characterize variability in the age at which autistic individuals attain key
developmental milestones based on co-occurring intellectual disability (ID), presence of
a rare disruptive genetic variant associated with neurodevelopmental disorders (NDD),
age at autism diagnosis, and research cohort membership.

DESIGN The study team harmonized data from 4 cross-sectional autism cohorts: the Autism
Genetics Research Exchange (n = 3284; 1997-2015), The Autism Simplex Collection (n = 694;
2008-2011), the Simons Simplex Collection (n = 2753; 2008-2011), and the Simons
Foundation Powering Autism Research for Knowledge (n = 10 367; 2016-present). The last
sample further included 4145 siblings without an autism diagnosis or ID.

PARTICIPANTS Convenience sample of 21 243 autistic individuals or their siblings without

an autism diagnosis aged 4 to 17 years.

MAIN OUTCOMES AND MEASURES Parents reported ages at which participants attained key
milestones including smiling, sitting upright, crawling, walking, spoon-feeding self, speaking
words, speaking phrases, and acquiring bladder and bowel control. A total of 5295 autistic
individuals, and their biological parents, were genetically characterized to identify de novo
variants in NDD-associated genes. The study team conducted time-to-event analyses to
estimate and compare percentiles in time with milestone attainment across autistic
individuals, subgroups of autistic individuals, and the sibling sample.

RESULTS Seventeen thousand ninety-eight autistic individuals (mean age, 9.15 years;
80.8% male) compared with 4145 siblings without autism or ID (mean age, 10.2 years;
50.2% female) showed delays in milestone attainment, with median (IQR) delays ranging
from 0.7 (0.3-1.6) to 19.7 (11.4-32.2) months. More severe and more variable delays in autism
were associated with the presence of co-occurring ID, carrying an NDD-associated rare
genetic variant, and being diagnosed with autism by age 5 years. More severe and more
variable delays were also associated with membership in earlier study cohorts, consistent
with autism’s diagnostic and ascertainment expansion over the last 30 years.

CONCLUSIONS AND RELEVANCE As the largest summary to date of developmental milestone

attainment in autism, to our knowledge, this study demonstrates substantial developmental
variability across different conditions and provides important context for understanding the
phenotypic and etiological heterogeneity of autism.

Author Affiliations: Author

affiliations are listed at the end of this
article.
Corresponding Author: Susan S.
Kuo, PhD, Broad Institute of MIT
and Harvard, 75 Ames St, Cambridge,
JAMA Pediatr. 2022;176(9):915-923. doi:10.1001/jamapediatrics.2022.2423 MA 02142 (susankuo@
Published online July 18, 2022. Last corrected on September 6, 2022. broadinstitute.org).

(Reprinted) 915

Downloaded From: https://jamanetwork.com/ by Fabricia Signorelli on 09/24/2022

 Research Original Investigation Developmental Variability in Autism Across 17 000 Autistic Individuals and 4000 Siblings Without an Autism Diagnosis
D
elays in milestone attainment (eg, delays in first words
or phrases) are often the first signs of developmental
differences in children who are later diagnosed with
autism.1 The extensive variability in developmental progres-
sion preceding an autism diagnosis2,3 suggests that develop-
mental milestone timing may capture an important facet of
autism heterogeneity.4 While many reports have provided es-
timates of milestone timing averages (central tendency), few
have characterized the full range of variability in milestone
attainment across multiple developmental domains.5-8 By
aggregating several large autism cohorts, we increase previ-
ous sample sizes by approximately an order of magnitude. In
this study, we aimed to estimate developmental milestone
averages and distributions in more than 17 000 autistic indi-
viduals, as well as in more than 4000 siblings of autistic chil-
dren who do not have an autism diagnosis themselves. We fur-
ther characterized milestone progression in several subgroups
of autistic individuals: autistic male individuals and autistic
female individuals, autistic individuals with co-occurring
intellectual disability (ID), autistic individuals with a genetic
variant associated with neurodevelopmental disorders
(NDD), autistic individuals diagnosed early (before age 5 years),
and autistic individuals diagnosed late (after age 10 years).
We also examined milestone attainment differences by re-
search cohort.
Many of these groups of autistic individuals have el-
evated rates of co-occurring ID. Co-occurring ID is associated
with global developmental delays in and outside of autism, and
is present in approximately 17% to 39% of autistic individuals
in the US.9 It is more common in autistic individuals who carry
disruptive, typically de novo, genetic variants in genes asso-
ciated with NDD.10-12 ID is also more common in female indi-
viduals diagnosed with autism,13,14 which could reflect sex dif-
ferences in both ascertainment and autism presentation.15,16
As ID, with or without autism, is associated with delays in at-
taining key milestones, groups of autistic individuals with
higher rates of co-occurring ID often attain milestones
later.12,17,18 The large amount of data aggregated here permits
full characterization, and comparison, of milestone attain-
ment across these groups of autistic individuals with ID.
Most averages in autism presentation are sensitive to
time and place of diagnosis and evolve with ascertainment
trends. On average, over the past decade and across 40 coun-
tries, children are diagnosed with autism around age 5 years.19
Autistic individuals diagnosed earlier in the US are more
likely to have co-occurring ID or global developmental delay.20
However, even as rates of ID in autism have declined,21,22 au-
tistic individuals are being diagnosed earlier on average than
in previous years.23,24 Thus, controlling for age at study en-
try, recent cohorts may include higher proportions of autistic
individuals who show milder developmental delays. To our
knowledge, no studies to date have compared milestone at-
tainment across autism cohorts ascertained across time.
Here, we explore these associations across cohorts ascer-
tained at different start dates over 25 years (1997 to present)
to clarify the scope and stability of developmental heteroge-
neity in autism that is being captured in prominent research
cohorts.
916 JAMA Pediatrics September 2022 Volume 176, Number 9 (Reprinted)
Downloaded From: https://jamanetwork.com/ by Fabricia Signorelli on 09/24/2022
Key Points
Question When do autistic individuals, on average, attain key
developmental milestones?
Findings Using retrospective, parent-reported data from
17 098 autistic individuals in this cross-sectional study, a found
substantial variability in average developmental milestone
acquisition was found. Average delays increased with co-occurring
intellectual disability, presence of a genetic variant associated with
neurodevelopmental disorders, earlier autism diagnosis, and
participation in older autism cohorts, and average delays also were
larger for later milestones (eg, phrase speech, bowel control) than
earlier milestones (eg, smiling, sitting).
Meaning These findings show that developmental milestone
progress in autism varies substantially under different conditions,
including presence of intellectual disability, genetic testing results,
diagnosis timing, and study cohort membership.
Methods
Sample
We aggregated data from 17 098 autistic individuals from 4 mul-
tisite cohorts (eTable 1 in the Supplement): the Autism Genetic
Resource Exchange (AGRE; n = 3284),25,26 the Autism Simplex
Collection (TASC; n = 694),27 the Simons Simplex Collection
(SSC; n = 2753),28 and the Simons Foundation Powering Au-
tism Research for Knowledge (SPARK; n = 10 367).29 The SPARK
cohort also collected milestone data from siblings without an
autism diagnosis; we therefore included 4145 siblings with-
out autism or ID from the SPARK cohort as a comparison group.
Autistic individuals and siblings without an autism diagnosis
were included if they had a measured value for at least 1 of the
developmental milestone variables (delineated below) and were
between the ages of 4 and 17 years (as those ages were in-
cluded in each of the 4 cohorts). Furthermore, 5295 autistic par-
ticipants (30.8% of the autistic sample) had available exome
sequencing data (including 43 from AGRE, 2048 from SSC, and
3204 from SPARK). Given that the milestones we aimed to char-
acterize are, on average, attained in the general population by
4 years of age,30 the participants had aged past the normative
age period for attaining the milestones of interest. The final
sample included autistic participants who were on average 9.15
years old and siblings who were on average 10.20 years old
(Table). Autistic participants were also 80% male, similar to
population-level sex ratios for autism,31 whereas male and fe-
male siblings were evenly represented. Written informed con-
sent and assent was provided for all participants in each co-
hort. The research specific to this study was approved by the
Partners Healthcare Institutional Review Board and was re-
ported following Strengthening the Reporting of Observa-
tional Studies in Epidemiology (STROBE) reporting guidelines.32
Procedures
At the time the participants were enrolled in each of the above
studies, parents retrospectively reported the age at which par-
ticipants attained the following developmental milestones
(1) smiling, (2) gross motor skills (including sitting upright,
jamapediatrics.com
 Developmental Variability in Autism Across 17 000 Autistic Individuals and 4000 Siblings Without an Autism Diagnosis Original Investigation Research

crawling, and walking), (3) self-help skills (spoon-feeding self),

Table. Sample Characteristics
(4) expressive language skills (speaking words and phrases),
and (5) toileting (acquiring bladder control and acquiring bowel Autistic Siblings without
individuals an autism diagnosis
control). Age at milestone attainment was coded in months. (n = 17 098), (n = 4145),
Characteristic No. (%) No. (%)
Different milestones had different sample sizes as not all mile-
Age, mean (SD), y 9.15 (3.6) 10.2 (3.81)
stones were assessed by each study cohort. A full description
of the sample size and assessment approach for each mile- Sex

stone can be found in eTable 2 in the Supplement. Analyses Male 13 816 (80.8) 2065 (49.8)
for each milestone were independent and not affected by dif- Female 3282 (19.2) 2080 (50.2)
fering sample sizes across milestones. Co-occurring intellectual 9691
disability
In the SSC, AGRE, and TASC cohorts, ID was defined as full-
Intellectual disability 3665 (37.8)
scale or nonverbal IQ score under 70. SPARK participants were
No intellectual disability 6026 (62.2)
also classified as having ID if they were reported by parents to
have an estimated IQ score under 70, ID, significant cognitive Genetic etiology 5295

impairment, global developmental delay, or borderline intel- Associated rare variant

lectual functioning. Presence of an NDD-associated genetic With NDD–associated 352 (6.6)
NA
rare variant
variant was defined to include de novo protein truncating vari-
Without NDD–associated 4943 (93.4)
ants, copy number variants, and missense variants with a rare variant
Missense PolyPhen Constraint score33 greater than 2 (indicat- Age at autism diagnosis, y 10 367
ing that it is highly disruptive, on par with protein truncating By age 5 6859 (66.2)
variants) in 1 of 373 genes associated with NDD at an approxi- At ages 5-9 2797 (27.0)
mately genome-wide significant level through a recent exome
After age 10 711 (6.9)
sequencing study.34
Abbreviations: NA, not applicable; NDD, neurodevelopmental disorders.

Statistical Analyses
Using observation-level data from each cohort, we modeled
milestone attainment using a time-to-event approach with
interval-censored outcomes,35 including (1) exact time-to-
event, where the milestone was attained at a known age, coded
[t, t); (2) right-censored data, where the milestone was not yet
attained by the age at assessment, t, coded [t, +⬁); (3) left-
censored data, where the milestone was attained by the age
of assessment, t, but the age at attainment was not known,
coded (–⬁, t]; and (4) interval-censored data, where the mile-
stone was attained after a specific age, t1, before the age at as-
sessment, t2, coded (t1, t2). Time-to-event analysis methods
are based on survival analysis methods; similar to how sur-
vival analyses estimate the time to a loss event in a popula-
tion, time-to-event analyses estimate the time to a gain event
population. We fitted parametric models in the overall au-
tism sample based on Weibull, gamma, and log-normal distri-
butions, and selected the model with the lowest Akaike infor-
mation criterion (AIC).36,37 Given that the AIC was consistently
lowest for the log-normal distribution across milestones, we
selected the log-normal hazard function for all subsequent
analyses.
We fitted time-to-event models in the overall autistic
sample and the sibling sample. We then conducted time-to-
event models by autistic subgroups based on (1) sex and ID,
(2) age at diagnosis, (3) genetic findings, and (4) study cohort.
We conducted these analyses 1 subgroup variable at a time,
such that the sample sizes vary between analyses in a way that
reflects which cohort(s) the subgroup variable was measured
in. We used these time-to-event models to identify the 5th,
25th, 45th, 50th (median), 55th, 75th, and 95th percentiles of
time to milestone attainment, equivalent to the ages at which
different proportions of the sample (ie, 5%, 25%, etc) at-
tained each milestone.
We conducted conservative 2-sided log-rank 2-sample per-
mutation tests (with 100 permutations) to test pairwise dif-
ferences between subgroup time-to-event distributions.38,39
We used false discovery rate (FDR)40 to correct for multiple
comparisons in each set of analyses (eg, for the sex and ID
analysis, given 4 subgroups of autistic individuals and 1 sib-
ling subgroup, we adjusted P values for 10 pairwise subgroup
comparisons per milestone, across 9 milestones, for a total of
90 pairwise comparisons). FDR-corrected P values were con-
sidered significant at P < .05.
Results
Overall Comparisons With Siblings
Compared with 4145 siblings without an autism diagnosis or
ID (mean age, 10.2 years; 50.2% female), 17 098 autistic indi-
viduals (mean age, 9.15 years; 80.8% male) showed delays in
milestone attainment, with median (IQR) delays ranging from
0.7 (0.3-1.6) to 19.7 (11.4-32.2) months. As expected, autistic
individuals showed substantial delays in attaining mile-
stones compared with their siblings without an autism diag-
nosis or ID, whose milestone attainment was generally con-
sistent with and sometimes slightly earlier than general
population norms.41-43 Of the autistic participants who pro-
vided data for phrase speech, bladder control, and bowel con-
trol, 502 participants (4.2%) had not attained any of these mile-
stones. eTable 3 in the Supplement details percentiles for
milestone attainment in autistic individuals, subgroups of au-
tistic individuals, siblings without an autism diagnosis or ID,
and a comparison set of general population estimates.41-43
eTable 4 in the Supplement presents pairwise comparisons of
subgroup time-to-event distributions.

jamapediatrics.com (Reprinted) JAMA Pediatrics September 2022 Volume 176, Number 9 917

Downloaded From: https://jamanetwork.com/ by Fabricia Signorelli on 09/24/2022

 Research Original Investigation
Figure 1. Developmental Milestones Attained by Autistic Individuals, Grouped by Cohort
Acquiring bowel
control
Speaking
phrases
Developmental milestone
Speaking
words
Walking
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 78 81 84 87 90 93 96
SPARK siblings without an autism diagnosis (n = 4145); SPARK autistic
individuals (n = 10 367); Simons Simplex Collection (n = 2753); The Autism
Simplex Collection (n = 694); and Autism Genetics Resource Exchange
Cohort
Autistic individuals were ascertained from as early as 1997 in
the AGRE cohort, 2008 in the TASC and SSC cohorts, and 2016
in the SPARK cohort. Compared with autistic individuals from
the most recent cohort, autistic individuals from earlier co-
horts showed greater and more variable delays for many mile-
stones, including speaking words, speaking phrases, and ac-
quiring bowel control (Figure 1 and eTable 5 in the Supplement).
Specifically, all time-to-event distributions differed signifi-
cantly, following multiple testing correction, between co-
horts except between AGRE vs SSC and TASC vs SPARK for
walking, SSC vs SPARK for speaking phrases, and TASC vs
SPARK for acquiring bladder control. All significant pairwise
cohort comparisons passed FDR-corrected P < .01, as de-
tailed in eTable 6 in the Supplement.
Co-occurring ID
As anticipated, autistic individuals with ID showed substan-
tial additional delays in milestone attainment (Figure 2; and
eTable 7 in the Supplement). We observed a gradient of sever-
ity and variability in developmental delays based on sex and
co-occurring ID, with male individuals with ID showing the
largest and most variable delays, followed by female individu-
als with ID, then male individuals without ID, then female in-
dividuals without ID, then siblings (eTable 8 in the Supple-
ment). These gaps were already noticeable for the most early,
basic milestones, and widened further as individuals pro-
gressed to later milestones. At the median ages for attaining
milestones, these delays spanned 1 to 2 years for expressive
language skills, such as speaking words and speaking phrases,
918 JAMA Pediatrics September 2022 Volume 176, Number 9 (Reprinted)
Downloaded From: https://jamanetwork.com/ by Fabricia Signorelli on 09/24/2022
Developmental Variability in Autism Across 17 000 Autistic Individuals and 4000 Siblings Without an Autism Diagnosis
Cohort
SPARK siblings without autism (2016-present)
SPARK individuals with autism (2016-present)
Simons Simplex Collection (2008-2011)
The Autism Simplex Collection (2008-2011)
25 50 75
Proportion attained, % Autism Genetics Resource Exchange (1997-2015)
Age, mo
(n = 3284). See eTable 3 in the Supplement for milestone attainment
percentiles and eTable 4 in the Supplement for pairwise subgroup comparisons
of survival distributions.
and for toileting skills, such as acquiring bladder and acquir-
ing bowel control. Notably, for individuals with co-occurring
ID, the 95th percentile extended beyond age 11 years, or with
a delay of more than 7 years, for speaking phrases, acquiring
bladder control, and acquiring bowel control.
The study team further observed sex differences in individu-
als with ID (eTable 8 in the Supplement) but not in individuals
without ID for earlier milestones (ie, smiling, sitting upright,
crawling, and acquiring bladder control). In contrast, we ob-
served the opposite trend for later milestones (ie, spoon-
feeding self, speaking words, speaking phrases, acquiring blad-
der control, and acquiring bowel control), finding sex differences
in individuals without ID but not in individuals with ID.
Genetic Etiology
Approximately 352 of the genetically characterized autistic in-
dividuals (6.6%) harbored a genetic variant that arose de novo
and disrupted the coding sequence of an NDD-associated gene.
Carrying an NDD-associated variant substantially increased de-
lays across all milestones (Figure 3 and eTable 9 in the Supple-
ment) except for smiling, for which autistic individuals with
a known variant did not differ from those without a known rare
variant (eTable 10 in the Supplement).
Diagnosis Timing
Developmental delays were most substantial for autistic indi-
viduals diagnosed at or before age 5 years, followed by autis-
tic individuals diagnosed at ages 5 to 9 years, then by autistic
individuals diagnosed after age 10 years (Figure 4, eTable 11,
and eTable 12 in the Supplement). Autistic children diagnosed
jamapediatrics.com
 Figure 2. Developmental Milestones Attained by Autistic Individuals, Grouped by Sex and Intellectual Disability (ID)

Autism overall 9.6

Autistic males with ID 13.0 Acquiring bowel control
Autistic females with ID 13.6 16.5% not yet attained
Autistic males without ID
Autistic females without ID

jamapediatrics.com
Siblings without autistic or ID
Autism overall
Autistic males with ID 11.2
Autistic females with ID Acquiring bladder control
12.5
Autistic males without ID 8.3% not yet attained
Autistic females without ID
Siblings without autistic or ID
Autism overall 8.8
Autistic males with ID 13.5
Autistic females with ID Speaking phrases
13.1
Autistic males without ID 10.0% not yet attained
Autistic females without ID
Siblings without autistic or ID
Autism overall
Autistic males with ID
Autistic females with ID Speaking words
Autistic males without ID 3.8% not yet attained
Autistic females without ID
Siblings without autistic or ID
Autism overall
Autistic males with ID
Autistic females with ID Spoon-feeding self
2.7% not yet attained

Downloaded From: https://jamanetwork.com/ by Fabricia Signorelli on 09/24/2022

Autistic males without ID

Group
Autistic females without ID
Siblings without autistic or ID
Autism overall
Developmental milestone

Autistic males with ID

Autistic females with ID Walking
Autistic males without ID 0.1% not yet attained
Autistic females without ID
Siblings without autistic or ID
Autism overall
Autistic males with ID
Autistic females with ID Crawling
Autistic males without ID 1.1% not yet attained
Autistic females without ID
Siblings without autistic or ID
Autism overall
Autistic males with ID
Developmental Variability in Autism Across 17 000 Autistic Individuals and 4000 Siblings Without an Autism Diagnosis

Autistic females with ID Sitting upright

Autistic males without ID 0.0% not yet attained
Autistic females without ID
Siblings without autistic or ID 5 25 45 55 75 95
Autism overall Proportion attained, %
Autistic males with ID
Autistic females with ID Smiling
Autistic males without ID 0.0% not yet attained
Autistic females without ID
Siblings without autistic or ID
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 78 81 84 87 90 93 96 99 102 105 108
Age, mo

Developmental milestones not yet attained for overall autism sample. Siblings without an autism diagnosis or ID an autism diagnosis or ID show comparable developmental milestone attainment compared with the general
(n = 4145); autistic males with ID (n = 2875); autistic females with ID (n = 790); autistic males without ID population.41-43 See eTable 7 in the Supplement for milestone attainment percentiles and eTable 8 in the
(n = 5000); autistic females without ID (n = 1026); and autistic individuals overall (n = 17 098). Siblings without Supplement for pairwise subgroup comparisons of survival distributions.

(Reprinted) JAMA Pediatrics September 2022 Volume 176, Number 9

Original Investigation Research

919
 Research Original Investigation Developmental Variability in Autism Across 17 000 Autistic Individuals and 4000 Siblings Without an Autism Diagnosis

Figure 3. Developmental Milestones Attained by Autistic Individuals, Grouped by Genetic Etiology

Developmental milestone
Smiling
(1.6-4.0)
Speaking phrases
(22.0-48.9) 25 50 75
Walking
(11.1-15.8) Proportion attained, %

Sitting upright
(4.9-7.7)
Acquiring bladder control
Spoon-feeding self (34.7-58.1)
(12.2-26.4)
Autism without
NDD-associated
Crawling
rare genetic variant
(6.5-11.7)
Acquiring bowel control
Speaking words (37.2-64.0)
(12.4-28.8)

Speaking words
(13.8-37.1) Acquiring bowel control
(39.8-81.7)
Crawling
(7.9-4.6)
Autism with
NDD-associated Spoon-feeding
rare genetic variant self
(14.1-37.5) Acquiring bladder control
(37.3-74.6)
Sitting upright
(5.7-9.8)

Walking
(13.0-20.8)
Speaking phrases
Smiling (25.1-61.0)
(1.6-4.3)

0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 78 81 84 87 90
Age, mo

Autism without neurodevelopmental disorder (NDD)–associated rare genetic cohorts.. See eTable 9 in the Supplement for milestone attainment percentiles
variant (n = 4943); and autistic with NDD-associated rare genetic variant and eTable 10 in the Supplement for pairwise subgroup comparisons of survival
(n = 352). Participants from AGRE, SPARK, and Simons Simplex Collection distributions.

by age 5 years were significantly more likely to have co-
occurring ID than autistic children diagnosed at ages 5 to 9 years
(56.1% vs 23.7%; χ 21 = 428.998; P < .001) or autistic children
diagnosed after age 10 years (56.1% vs 25.7%; χ 21 = 172.985;
P < .001). The latter 2 groups did not differ significantly from
each other (23.7% vs 25.7%; χ 21 = 0.675; P = .41). We found no
differences between those diagnosed at ages 5 to 9 years and
those diagnosed after age 10 years for early milestones includ-
ing smiling, sitting upright, and crawling. Furthermore, for
crawling and walking, cases diagnosed by age 5 years did not
differ significantly from those diagnosed at ages 5 to 9 years.
Discussion
Early milestone attainment is highly variable among autistic
individuals. More severe and more variable delays in autism
were associated with the presence of co-occurring ID, carry-
ing an NDD-associated genetic variant, and being diagnosed
with autism by age 5 years. More severe and more variable de-
lays were also associated with membership in earlier study co-
horts, consistent with autism’s diagnostic and ascertainment
expansion over the last 30 years. In light of the observed co-
hort effects, our findings should be taken in the context of con-
stantly evolving conceptualizations of how core autism symp-
toms manifest. Acknowledging that no single study to date,
no matter how large, can be fully or stably representative of
autism, our findings highlight the extent to which estimates
of milestone attainment, and many other phenotypic aver-
ages in autism, are influenced by diagnosis timing and
ascertainment.44-47
Co-occurring ID substantially extends the timeline of de-
velopmental milestone attainment in autism. Age at walking
was an exception, however, as autistic individuals with and
without ID did not differ in this milestone. These findings
build upon previous work suggesting that for some autistic
children (ie, many of whom acquire early milestones as ex-
pected), ID may be a secondary consequence of the autism.
Thus, differences in attainment of basic milestones like age at
walking may not differentiate between autistic individuals with
or without ID to the same extent within autism cohorts (com-
pared with non–autism cohorts), and are more likely to signal

920 JAMA Pediatrics September 2022 Volume 176, Number 9 (Reprinted) jamapediatrics.com

Downloaded From: https://jamanetwork.com/ by Fabricia Signorelli on 09/24/2022

 Figure 4. Developmental Milestones Attained by Autistic Individuals, Grouped by Age at Autism Diagnosis

Autism diagnosis by age 5 y Autism diagnosis at ages 5-9 y Autism diagnosis after age 10 y Siblings without autism or ID
132 132 132 132
NDD-associated rare variant rate: 5.9% NDD-associated rare variant rate: 7.4% NDD-associated rare variant rate: 5.1% Proportion
126 Intellectual disability: 56.1% 126 Intellectual disability: 25.7% 126 Intellectual disability: 23.7% 126
attained, %

jamapediatrics.com
120 120 120 120 95

114 114 114 114 75

50
108 108 108 108
25
102 102 102 102 5
96 96 96 96

90 90 90 90

84 84 84 84

78 78 78 78

72 72 72 72

66 66 66 66

Age, mo
60 60 60 60

54 54 54 54

48 48 48 48

Downloaded From: https://jamanetwork.com/ by Fabricia Signorelli on 09/24/2022

42 42 42 42

36 36 36 36

30 30 30 30

24 24 24 24

18 18 18 18

12 12 12 12

6 6 6 6

0 0 0 0
g t g g l f s es l l g t g g lf s es l l g t g g lf s es l l ht g lf s es l l
gh in rd ro ro gh in rd ro ro gh in rd ro ro ng ng rd ro ro
Developmental Variability in Autism Across 17 000 Autistic Individuals and 4000 Siblings Without an Autism Diagnosis

i lin ri lin k se as nt nt i lin ri lin k se as nt nt i lin ri lin k se as nt nt ili rig lin ki se as nt nt

Sm up raw Wal i ng wo phr c o co Sm up raw Wal i ng wo phr c o c o Sm up raw Wal i ng wo phr c o c o Sm up raw Wal ing wo phr c o c o
g C d r l C d r l C d r l C d r l
in ee ki
ng ng
de we ng ee ki
ng ng
de we ng ee ki
ng ng
de we ng ee ki
ng ng
de we
tt -f ea ki ad bo tti -f ea ki ad bo tti -f ea ki ad bo tti -f ea ki ad bo
Si on Sp p ea bl g Si on Sp p ea bl g Si on Sp p ea bl g Si on Sp pea bl g
o S n o S n o S n o S n
Sp ng iri Sp ng iri Sp ng iri Sp ng iri
iri qu iri qu iri qu iri qu
qu Ac qu Ac qu Ac qu Ac
Ac Ac Ac Ac
Developmental milestone Developmental milestone Developmental milestone Developmental milestone

Autism diagnosed by age 5 years (n = 6859); autism diagnosed at age 5 to 9 years (n = 2797); autism diagnosed eTable 12 in the Supplement for pairwise subgroup comparisons of survival distributions. NDD indicates
after age 10 years (n = 711); and siblings without an autism diagnosis or intellectual disability (ID) (n = 4145). neurodevelopmental disorders.
Participants from SPARK cohort. See eTable 11 in the Supplement for milestone attainment percentiles and

(Reprinted) JAMA Pediatrics September 2022 Volume 176, Number 9

Original Investigation Research

921
 Research Original Investigation Developmental Variability in Autism Across 17 000 Autistic Individuals and 4000 Siblings Without an Autism Diagnosis

the presence of a rare genetic variant in autism.6,48,49 The un-
precedentedly large scale of this study also enabled us to ex-
amine developmental variability in individuals who carry an
NDD-associated rare genetic variant. We used a strict thresh-
old to define genetic variants. Given that most genetic variants
are not captured in this category (such as common variants
that each contribute less to overall likelihood of an autism di-
agnosis than rare variants),50,51 future research will be able to
consider a broader set of genetic variants.
Our study bears several advantages over prior studies. We
harmonized 4 cohorts to evaluate multiple developmental mile-
stones across diverse behavioral domains, and included a large
comparison sample of siblings without an autism diagnosis or
ID. Participants were sampled from multiple developmental
periods spanning childhood to late adolescence, enabling us
to measure a wide range of possible ages for attaining develop-
mental milestones. Given that population health registries to date
have recorded relatively limited information on developmental
milestones, this study is the most comprehensive analysis at
this scale of developmental milestone attainment in autism.
Limitations
Our findings should be considered in light of certain limita-
tions. Variability in developmental milestone attainment is
likely attributable not only to autism but to contextual fac-
tors, such as cultural practices, particularly given that most of
the sample was based in the US. Our measures of develop-
mental progress were retrospective and based on parent re-
call and would be bolstered by converging evidence from lon-
gitudinal measures, such as prospective clinician records.
Siblings without an autism diagnosis showed slightly earlier
ages at attainment for some milestones compared with the gen-
eral population, which may reflect differences between chil-
dren without an autism diagnosis ascertained based on hav-
ing an autistic sibling vs children without an autism diagnosis
ascertained in the general population. Similarly, siblings with-
out an autism diagnosis were from 1 cohort and may not be rep-
resentative of siblings without an autism diagnosis from other
cohorts. The cohort effects suggest that the autism sample is
not, and cannot be, representative of all autistic individuals.
This limitation is not unique to our study design and remains
challenging for all autism studies.
Conclusion
As the largest summary to date of developmental milestone at-
tainment in autism, this study highlights the extensive develop-
mental variability within and across different contexts of autism.
Our work emphasizes the utility of capturing not only average
developmental progress, but also variability in developmental
progress, to elucidate the causes and courses of autism.

ARTICLE INFORMATION
Accepted for Publication: May 14, 2022.
Published Online: July 18, 2022.
doi:10.1001/jamapediatrics.2022.2423
Open Access: This is an open access article
distributed under the terms of the CC-BY License.
© 2022 Kuo SS et al. JAMA Pediatrics.
Correction: This article was corrected on
September 6, 2022, to update the article’s open
access status.
Author Affiliations: Stanley Center for Psychiatric
Genetics, Broad Institute of MIT and Harvard,
Cambridge, Massachusetts (Kuo, van der Merwe,
Fu, Carey, Talkowski, Robinson); Center for
Genomic Medicine, Massachusetts General
Hospital, Harvard Medical School, Boston (Kuo,
van der Merwe, Fu, Carey, Talkowski, Robinson);
Analytic and Translational Genetics Unit,
Massachusetts General Hospital, Harvard Medical
School, Boston (Carey); Department of Neurology,
Massachusetts General Hospital, Harvard Medical
School, Boston (Talkowski); Department of
Psychiatry, University of California, San Francisco
(Bishop); Department of Epidemiology,
Harvard T. H. Chan School of Public Health,
Boston, Massachusetts (Robinson); Department
of Psychiatry, Massachusetts General Hospital,
Harvard Medical School, Boston (Robinson).
Author Contributions: Drs Kuo and Robinson had
full access to all of the data in the study and take
responsibility for the integrity of the data and the
accuracy of the data analysis.
Concept and design: Kuo, Bishop, Robinson.
Acquisition, analysis, or interpretation of data: Kuo,
Fu, Talkowski, Bishop, Robinson.
Drafting of the manuscript: Kuo, Robinson.
Critical revision of the manuscript for important
intellectual content: All authors.
Statistical analysis: Kuo, Bishop, Robinson.
Obtained funding: van der Merwe, Fu, Talkowski,
Bishop, Robinson.
Administrative, technical, or material support:
van der Merwe, Fu, Carey.
Supervision: Bishop, Robinson.
Conflict of Interest Disclosures: Dr Fu reported
grants from Autism Speaks Postdoctoral Fellowship
during the conduct of the study. Dr Talkowski
reported grants from the National Institutes of
Health during the conduct of the study,
nonfinancial support from Illumina Inc, grants from
Microsoft Inc and Levo Therapeutics outside the
submitted work. Dr Bishop reported personal fees
from Western Psychological Services Royalties
for sales of the ADOS-2 and personal fees from
Janssen Paid consultant outside the submitted
work. No other disclosures were reported.
Funding/Support: This study was supported by
Simons Foundation Autism Research Initiative
(SSC 573206 and SPARK 606362), National
Institute of Mental Health (R01-MH111813 and
R01-MH115957), National Human Genome
Research Institute (R01-HG00895), Autism
Speaks, Brain and Behavior Foundation,
Autism Science Foundation, and The Hilibrand
Foundation (Autism Sisters Project).
Role of the Funder/Sponsor: The funders had no
role in the design and conduct of the study;
collection, management, analysis, and
interpretation of the data; preparation, review,
or approval of the manuscript; and decision to
submit the manuscript for publication.
Additional Contributions: We thank Alycia
Halladay, PhD, Autism Science Foundation, for her
helpful feedback on an early draft of the article.
REFERENCES
1. Jones EJ, Gliga T, Bedford R, Charman T,
Johnson MH. Developmental pathways to autism:
a review of prospective studies of infants at risk.
Neurosci Biobehav Rev. 2014;39(100):1-33.
doi:10.1016/j.neubiorev.2013.12.001
2. Billstedt E, Gillberg IC, Gillberg C. Autism after
adolescence: population-based 13- to 22-year
follow-up study of 120 individuals with autism
diagnosed in childhood. J Autism Dev Disord. 2005;
35(3):351-360. doi:10.1007/s10803-005-3302-5
3. Waizbard-Bartov E, Ferrer E, Young GS, et al.
Trajectories of autism symptom severity change
during early childhood. J Autism Dev Disord. 2021;51
(1):227-242. doi:10.1007/s10803-020-04526-z
4. Hirota T, Bishop S, Adachi M, et al. Utilization of
the Maternal and Child Health Handbook in early
identification of autism spectrum disorder and
other neurodevelopmental disorders. Autism Res.
2021;14(3):551-559. doi:10.1002/aur.2442
5. Reindal L, Nærland T, Weidle B, Lydersen S,
Andreassen OA, Sund AM. Age of first walking and
associations with symptom severity in children with
suspected or diagnosed autism spectrum disorder.
J Autism Dev Disord. 2020;50(9):3216-3232.
doi:10.1007/s10803-019-04112-y
6. Bishop SL, Thurm A, Farmer C, Lord C.
Autism spectrum disorder, intellectual disability,
and delayed walking. Pediatrics. 2016;137(3):
e20152959. doi:10.1542/peds.2015-2959
7. Øien RA, Schjølberg S, Volkmar FR, et al.
Clinical features of children with autism who passed
18-month screening. Pediatrics. 2018;141(6):
e20173596. doi:10.1542/peds.2017-3596

922 JAMA Pediatrics September 2022 Volume 176, Number 9 (Reprinted) jamapediatrics.com

Downloaded From: https://jamanetwork.com/ by Fabricia Signorelli on 09/24/2022

 Developmental Variability in Autism Across 17 000 Autistic Individuals and 4000 Siblings Without an Autism Diagnosis
8. Matson JL, Mahan S, Kozlowski AM,
Shoemaker M. Developmental milestones in
toddlers with autistic disorder, pervasive
developmental disorder—not otherwise specified
and atypical development. Dev Neurorehabil. 2010;
13(4):239-247. doi:10.3109/17518423.2010.481299
9. Maenner MJ, Shaw KA, Bakian AV, et al.
Prevalence and characteristics of autism spectrum
disorder among children aged 8 years—Autism and
Developmental Disabilities Monitoring Network,
11 Sites, United States, 2018. MMWR Surveill Summ.
2021;70(11):1-16. doi:10.15585/mmwr.ss7011a1
10. Kaplanis J, Samocha KE, Wiel L, et al;
Deciphering Developmental Disorders Study.
Evidence for 28 genetic disorders discovered by
combining healthcare and research data. Nature.
2020;586(7831):757-762. doi:10.1038/s41586-020-
2832-5
11. Iossifov I, O’Roak BJ, Sanders SJ, et al.
The contribution of de novo coding mutations
to autism spectrum disorder. Nature. 2014;515
(7526):216-221. doi:10.1038/nature13908
12. Satterstrom FK, Kosmicki JA, Wang J, et al;
Autism Sequencing Consortium; iPSYCH-Broad
Consortium. Large-scale exome sequencing study
implicates both developmental and functional
changes in the neurobiology of autism. Cell. 2020;
180(3):568-584.e23. doi:10.1016/j.cell.2019.12.036
13. Rødgaard EM, Jensen K, Miskowiak KW,
Mottron L. Autism comorbidities show elevated
female-to-male odds ratios and are associated with
the age of first autism diagnosis. Acta Psychiatr Scand.
2021;144(5):475-486. doi:10.1111/acps.13345
14. Christensen DL, Baio J, Van Naarden Braun K,
et al; Centers for Disease Control and Prevention
(CDC). Prevalence and characteristics of autism
spectrum disorder among children aged 8
years–Autism and Developmental Disabilities
Monitoring Network, 11 sites, United States,
2012. MMWR Surveill Summ. 2016;65(3):1-23.
doi:10.15585/mmwr.ss6503a1
15. Kaat AJ, Shui AM, Ghods SS, et al. Sex
differences in scores on standardized measures
of autism symptoms: a multisite integrative data
analysis. J Child Psychol Psychiatry. 2021;62(1):
97-106. doi:10.1111/jcpp.13242
16. Van Wijngaarden-Cremers PJ, van Eeten E,
Groen WB, Van Deurzen PA, Oosterling IJ,
Van der Gaag RJ. Gender and age differences in
the core triad of impairments in autism spectrum
disorders: a systematic review and meta-analysis.
J Autism Dev Disord. 2014;44(3):627-635.
doi:10.1007/s10803-013-1913-9
17. Wickstrom J, Farmer C, Green Snyder L, et al.
Patterns of delay in early gross motor and
expressive language milestone attainment in
probands with genetic conditions versus idiopathic
ASD from SFARI registries. J Child Psychol Psychiatry.
2021;62(11):1297-1307. doi:10.1111/jcpp.13492
18. Bishop SL, Farmer C, Bal V, et al. Identification
of developmental and behavioral markers
associated with genetic abnormalities in autism
spectrum disorder. Am J Psychiatry. 2017;174(6):
576-585. doi:10.1176/appi.ajp.2017.16101115
19. van ’t Hof M, Tisseur C, van Berckelear-Onnes I,
et al. Age at autism spectrum disorder diagnosis:
a systematic review and meta-analysis from 2012
to 2019. Autism. 2021;25(4):862-873. doi:10.1177/
1362361320971107
20. Zuckerman KE, Lindly OJ, Sinche BK.
Parental concerns, provider response, and
timeliness of autism spectrum disorder diagnosis.
J Pediatr. 2015;166(6):1431-9.e1. doi:10.1016/j.jpeds.
2015.03.007
jamapediatrics.com
Downloaded From: https://jamanetwork.com/ by Fabricia Signorelli on 09/24/2022
21. Idring S, Lundberg M, Sturm H, et al. Changes
in prevalence of autism spectrum disorders in
2001-2011: findings from the Stockholm youth
cohort. J Autism Dev Disord. 2015;45(6):1766-1773.
doi:10.1007/s10803-014-2336-y
22. Polyak A, Kubina RM, Girirajan S. Comorbidity
of intellectual disability confounds ascertainment
of autism: implications for genetic diagnosis. Am J
Med Genet B Neuropsychiatr Genet. 2015;168(7):
600-608. doi:10.1002/ajmg.b.32338
23. Keyes KM, Susser E, Cheslack-Postava K,
Fountain C, Liu K, Bearman PS. Cohort effects
explain the increase in autism diagnosis among
children born from 1992 to 2003 in California. Int J
Epidemiol. 2012;41(2):495-503. doi:10.1093/ije/
dyr193
24. Myers SM, Voigt RG, Colligan RC, et al.
Autism spectrum disorder: incidence and time
trends over two decades in a population-based
birth cohort. J Autism Dev Disord. 2019;49(4):1455-
1474. doi:10.1007/s10803-018-3834-0
25. Geschwind DH, Sowinski J, Lord C, et al;
AGRE Steering Committee. The autism genetic
resource exchange: a resource for the study of
autism and related neuropsychiatric conditions.
Am J Hum Genet. 2001;69(2):463-466.
doi:10.1086/321292
26. Lajonchere CM; AGRE Consortium. Changing
the landscape of autism research: the autism
genetic resource exchange. Neuron. 2010;68(2):
187-191. doi:10.1016/j.neuron.2010.10.009
27. Buxbaum JD, Bolshakova N, Brownfeld JM,
et al. The Autism Simplex Collection: an
international, expertly phenotyped autism sample
for genetic and phenotypic analyses. Mol Autism.
2014;5(1):34. doi:10.1186/2040-2392-5-34
28. Fischbach GD, Lord C. The Simons Simplex
Collection: a resource for identification of autism
genetic risk factors. Neuron. 2010;68(2):192-195.
doi:10.1016/j.neuron.2010.10.006
29. SPARK Consortium. SPARK: A US cohort
of 50,000 families to accelerate autism research.
Neuron. 2018;97(3):488-493. doi:10.1016/j.neuron.
2018.01.015
30. Sheldrick RC, Schlichting LE, Berger B, et al.
Establishing new norms for developmental
milestones. Pediatrics. 2019;144(6):e20190374.
doi:10.1542/peds.2019-0374
31. Loomes R, Hull L, Mandy WPL. What Is the
male-to-female ratio in autism spectrum disorder?
a systematic review and meta-analysis. J Am Acad
Child Adolesc Psychiatry. 2017;56(6):466-474.
doi:10.1016/j.jaac.2017.03.013
32. von Elm E, Altman DG, Egger M, Pocock SJ,
Gøtzsche PC, Vandenbroucke JP; STROBE Initiative.
The Strengthening the Reporting of Observational
Studies in Epidemiology (STROBE) statement:
guidelines for reporting observational studies. Lancet.
2007;370(9596):1453-1457. doi:10.1016/S0140-6736
(07)61602-X
33. Samocha KE, Kosmicki JA, Karczewski KJ, et al.
Regional missense constraint improves variant
deleteriousness prediction. bioRxiv. 2017:148353.
doi:10.1101/148353
34. Fu JM, Satterstrom FK, Peng M, et al.
Rare coding variation illuminates the allelic
architecture, risk genes, cellular expression
patterns, and phenotypic context of autism. medRxiv.
2021:2021.12.20.21267194. doi:10.1101/2021.12.20.
21267194
35. Anderson-Bergman C. icenReg: regression
models for interval censored data in R. J Stat Softw.
2017;81(1):1-23. doi:10.18637/jss.v081.i12
(Reprinted) JAMA Pediatrics September 2022 Volume 176, Number 9
Original Investigation Research
36. Akaike H. A new look at the statistical model
identification. IEEE Trans Automat Contr. 1974;19
(6):716-723. doi:10.1109/TAC.1974.1100705
37. Bradburn MJ, Clark TG, Love SB, Altman DG.
Survival analysis part III: multivariate data
analysis—choosing a model and assessing its
adequacy and fit. Br J Cancer. 2003;89(4):605-611.
doi:10.1038/sj.bjc.6601120
38. Fay MP, Shaw PA. Exact and asymptotic
weighted logrank tests for interval censored data:
the interval R package. J Stat Softw. 2010;36(2):i02.
doi:10.18637/jss.v036.i02
39. Finkelstein DM. A proportional hazards model
for interval-censored failure time data. Biometrics.
1986;42(4):845-854. doi:10.2307/2530698
40. Benjamini Y, Yekutieli D. The control of the
false discovery rate in multiple testing under
dependency. Ann Stat. 2001;29(4):1165-1188.
doi:10.1214/aos/1013699998
41. Schum TR, Kolb TM, McAuliffe TL, Simms MD,
Underhill RL, Lewis M. Sequential acquisition of
toilet-training skills: a descriptive study of gender
and age differences in normal children. Pediatrics.
2002;109(3):E48. doi:10.1542/peds.109.3.e48
42. Frankenburg WK, Dodds J, Archer P, Shapiro H,
Bresnick B. The Denver II: a major revision and
restandardization of the Denver Developmental
Screening Test. Pediatrics. 1992;89(1):91-97.
doi:10.1542/peds.89.1.91
43. WHO Multicentre Growth Reference Study
Group. WHO Motor Development Study: windows
of achievement for six gross motor development
milestones. Acta Paediatr Suppl. 2006;450:86-95.
doi:10.1111/j.1651-2227.2006.tb02379.x
44. Russell G, Mandy W, Elliott D, White R,
Pittwood T, Ford T. Selection bias on intellectual
ability in autism research: a cross-sectional review
and meta-analysis. Mol Autism. 2019;10:9.
doi:10.1186/s13229-019-0260-x
45. Howe YJ, Yatchmink Y, Viscidi EW, Morrow EM.
Ascertainment and gender in autism spectrum
disorders. J Am Acad Child Adolesc Psychiatry.
2014;53(6):698-700. doi:10.1016/j.jaac.2014.04.003
46. Micheletti M, McCracken C, Constantino JN,
Mandell D, Jones W, Klin A. Research review:
outcomes of 24- to 36-month-old children with
autism spectrum disorder vary by ascertainment
strategy: a systematic review and meta-analysis.
J Child Psychol Psychiatry. 2020;61(1):4-17.
doi:10.1111/jcpp.13057
47. Daniels AM, Mandell DS. Explaining differences
in age at autism spectrum disorder diagnosis:
a critical review. Autism. 2014;18(5):583-597.
doi:10.1177/1362361313480277
48. Havdahl A, Farmer C, Schjølberg S, et al.
Age of walking and intellectual ability in autism
spectrum disorder and other neurodevelopmental
disorders: a population-based study. J Child Psychol
Psychiatry. 2021;62(9):1070-1078. doi:10.1111/jcpp.
13369
49. Iverson JM. Developmental variability and
developmental cascades: lessons from motor and
language development in infancy. Curr Dir Psychol Sci.
2021;30(3):228-235. doi:10.1177/0963721421993822
50. Gaugler T, Klei L, Sanders SJ, et al. Most genetic
risk for autism resides with common variation. Nat
Genet. 2014;46(8):881-885. doi:10.1038/ng.3039
51. Klei L, McClain LL, Mahjani B, et al. How rare
and common risk variation jointly affect liability for
autism spectrum disorder. Mol Autism. 2021;12(1):66.
doi:10.1186/s13229-021-00466-2
923