Appearance of the Hymen at Birth and One Year of Age: A Longitudinal Study

(1)

Appearance

of

the

Hymen

at Birth

and

One

Year

of

Age:

A Longitudinal

Study

Abbey B. Benenson, MD

ABSTRACT. The increase in the number of prepubertal

girls who require evaluation of possible sexual abuse

creates a need for detailed information, not previously

available from cross-sectional studies, on the influence of

aging on the hymen’s appearance. This study was

under-taken to evaluate and document, using a longitudinal

design, changes in the hymen’s morphology in 62 girls

without a history of sexual abuse between birth and their

first birthday. Labial agglutination extensive enough to

obscure the inferior half of the hymen was observed in 5

girls (8%) at 1 year of age. Thirty-three (58%) of the

re-maining 57 infants experienced a marked decrease in the

amount of their hymenal tissue between birth and 1 year.

Significantly more infants at 1 year of age had a

crescen-tic configuration (0% vs 28%), and significantly fewer

had an external ridge (82% vs 14%) as compared to the

newborn period. An annular hymen with a central or

ventrally displaced opening progressed to a crescentic

hymen in 13 children by 1 year, 77% (10/13) of whom

were observed to have a notch (cleft) at the 12 o’clock

position on the earlier study. A superior notch appeared

for the first time in 9 girls. Lateral notches resolved in 5

cases and persisted in 2. Inferior notches between the 4

and 8 o’clock positions were not observed at birth or 1

year. Hymenal tags resolved in 2 instances, persisted in

the same location in 2, and appeared for the first time in

4 cases. Three girls had a hymenal mound (bump) at 1

year, all of which could be traced back to a similar

find-ing at birth. No change in the number of infants with

longitudinal intravaginal ridges was observed.

Clini-cians should be aware of the influence of age and

chang-ing estrogen levels on the hymen’s morphology in order

to differentiate normal anatomical from posttraumatic or

infectious changes. Pediatrics 199391:820-825; hymen,

hymenal configuration, newborn, infant, sexual abuse.

ABBREVIATION. CL, confidence limits.

As the number of prepubertal children who

re-quire

evaluation for possible sexual abuse increases,

the need to develop normal standards for the

hy-men’s appearance becomes more acute.

Cross-sec-tional studies of nonabused, pnepubertal girls

pro-vide detailed descriptions of congenital findings,

hymenal opening sizes, and the presence of notches,

bumps, ridges, and bands on the hymen’s rim.13 The

From the Departments of Obstetrics and Gynecology and Pediatrics, Uni-versity of Texas Medical Branch, Galveston.

Received for publication Apr 15, 1992; accepted Jun 23, 1992.

Reprint requests to (A.B.B.) University of Texas Medical Branch, Depts of

Obstetrics and Gynecology, Section of Pediatric and Adolescent Gynecology, Galveston, TX 77555-0587.

effects of aging, combined with varying estrogen

1ev-els as the child advances in years, have not been

studied.

Investigations of the hymen in prepubentab girls

suggest that the anatomy of this structure varies with

age. Cross-sectional studies show that the hymenab

opening increases as the child ages.2’3 A recent study

of prepubertal girls from this institution observed

that the crescentic hymen occurs more frequently in

children 4 years old compared with younger

chil-dren and does not appear in neonates.4

Although cross-sectional studies provide a general

description of the external genitalia at different ages,

they do not outline the evolution of the hymen’s

appearance over time. This study was undertaken to

evaluate and document, with photographs, changes

in the hymen’s appearance in the first year of life in

a cohort of girls without a history of sexual abuse.

MATERIALS AND METHODS

All infants born between August 20, 1989, and November 13, 1989, at the University of Texas Medical Branch, whose external genitalia were photographed at birth under a previous protocol4

and who did not have a history of sexual abuse, were eligible to

participate in this longitudinal study. Of the 449 infants who qual-ified, a current address was available for 362 families, all of whom were contacted by mail requesting permission to photograph the child within 4 months of her first birthday. Parents of 57 (16%) infants agreed to participate. Details of the study were explained by phone and in person, outlining the need to develop a normative

database. An additional 5 infants who had been photographed at

6 weeks of age under a different protocol2 were recruited and included in the longitudinal study. All parents who participated

were questioned regarding a history of possible sexual abuse of their daughter. The medical records of each infant were also re-viewed to detect any previous examinations for possible assault. A prior history or suspicion of sexual abuse was not revealed by history or chart review in any of the 62 girls. Parents of children photographed at I year were reimbursed $10 for travel expenses.

With Institutional Review Board approval, each child was pho-tographed in the supine position with her legs extended upward.

The lower portion of her labia majora was grasped between the examiner’s thumb and index fingers and gently pulled outward (labial traction technique). Photographs were taken both at birth and again near the child’s first birthday (hereinafter referred to as 1-year-old photographs). When the vaginal opening was not vi-sualized at 1 year, the edges of the hymen were gently teased apart with a moistened Q-tip or fingertip. Photographs at birth and 1 year were taken by one of two examiners. One clinician (A.B.B.) examined 92% (57/62) of newborns and 68% (42/62) of infants at 1 year of age.

A hand-held Nikon F3 camera with a 105-mm micro-Nikkor

lens maintained fully extended with an attached 52.5-mm exten-sion ring and ring flash was used for all photographs. This re-sulted in I .2x magnification. An average of three slides per new-born and eight per 1-year-old were coded by study number and

placed in secure storage.

At the conclusion of the study, all newborn and 1-year-old slides were reviewed; findings reported in this paper reflect those

at Viet Nam:AAP Sponsored on September 1, 2020

www.aappublications.org/news

(2)

observed photographically. Characteristics recorded included

hy-menal configuration, vascular pattern, labial agglutination, and

the number and location of hymenal notches (clefts), bumps, tags, cysts, vestibular bands, and both external and intravaginal ridges as defined previously.24 Annular or crescentic hymens with a ventrally displaced orifice (microperforate) were further classified as sleeve configurations. Location was noted using the face of a clock with the infant in the supine position, with the 12 o’clock

position located ventrally under the urethra and the 6 o’clock

position dorsally near the fourchette. Longitudinal intravaginal

ridges, which were defined as ridges extending to or beyond the hymenal rim, were not identified at the 12 o’clock position because

of the location of the urethra at this position. A notch was defined as a division or split in the rim. Notches were not recorded in the fimbriated hymen because of its fringed nature nor between the U and I o’clock positions in a crescentic hymen where there is a normal absence of hymenal tissue. An external ridge appeared as a longitudinal ridge of the vestibular hymen from the rim to the

fossa naviculans or urethra. Vestibular bands were symmetrical bands which extended from the hymen to the vestibule. When an area was not visible photographically in all subjects, the number of subjects in whom the area could be seen is reported.

Measure-ments were not done because of redundancy of the tissue in all neonates and a large portion of 1-year-olds.

The prevalence of each characteristic among the 57 newborns

who were reexamined at 1 year of age was calculated. Prevalence values by characteristic at birth and I year old and 95% confidence limits around each prevalence value were determined using an algorithm developed by Rothman and Boice.5 If the confidence interval around the prevalence observed at one time period did not overlap the confidence interval around the prevalence ob-served at the other time period, the two prevalence values are

significantly different.

RESULTS

Fifty-seven infants of the 449 who participated in

the newborn study were photographed at I year of

age. Of these, 19 (33%) were white non-Hispanic, 14

(25%) were Hispanic, 23 (40%) were black, and I (2%)

was Asian. Compared with the racial distribution of

the original 419 infants with acceptable photographs

as newborns, more blacks (40% vs 26%) and fewer

whites (33% vs 45%) elected to participate at I year.

Three black and 2 Hispanic infants who had

photo-graphs taken between birth and 6 weeks of age were

also included, making a total of 62 infants in this

study. The mean age of subjects was 4.3 days at the

first set of photographs and 12.9 months at the

sec-ond. The inferior half of the hymen was obscured by

labial agglutination in 5 of the 62 (3 Hispanic, I black,

and I white non-Hispanic) 1-year-old infants.

Char-actenistics of the hymen are reported for the remain-ing 57 infants. Partial posterior labial agglutination

not obscuring the hymen was observed in 5% (3/57)

of these subjects at the time of the second set of

photographs.

The annular hymen was the most common

config-uration at both ages followed by the fimbniated in the

newborn

period and the crescentic at I year of age

(Table). Twenty-five of the 40 newborns with an

an-nular hymen and 4 of the 12 born with a fimbniated

hymen had persistence of the same configuration at I

year of age (Fig I,). All 4 newborns observed to have

a ventralby displaced orifice (sleeve configuration)

had persistence of this configuration at I year. A

change in configuration was observed in 24 girls.

None of the newborns were observed to have a

cres-centic configuration at birth (95% confidence limits

[CL] = 0, 6%). However, by I year of age,

signifi-cantly more of the same children were observed to

have crescentic hymens (28%, 95% CL = 10%, 41%).

An annular hymen with a central on ventral orifice

progressed to a crescentic hymen in 13 children by 1

year of age, 77% (10/13) of whom were observed to

have a notch at the 12 o’clock position on the earlier

photographs (Fig 2). Of the 12 fimbniated hymens

observed at birth, 5 progressed to annular and 3 to

crescentic (Fig 3). Two additional children who had a

central orifice at birth appeared to have a hymen

with a ventrally displaced orifice by I year. A

sep-tated hymen at birth appeared annular without a

septum at I year.

Hymenal tissue appeared redundant in all

neo-nates. A marked decrease in tissue resulting in a

membrane without folds was observed in 58% (33/

57) of infants at I year of age (Fig 3). Moreover, a

portion of the hymen, usually on the inferior half,

appeared translucent, revealing its thinness by the

second set of photographs in 30% (10/33) of those

with a stretched membrane.

Notches, observed in 38% of newborns and 29% of

1-yean-olds, most frequently appeared superiorly at

both ages. Ten newborns with a notch between the II

and I o’clock positions and 5 with one or two lateral

notches did not have this finding by the second set of

photographs (Fig 4). One infant had persistence of a

superior notch at I year. Nine children, who had not

had this finding previously, developed a notch

be-tween the II and I o’clock positions by the I-year

photographs. Lateral notches were observed in only

two I-yean-olds, both of whom had a similar finding

TABLE. Classification of Hymenal Findin gs at Birth an d One Year of Ag e (n = 57)

Newborn 1 Year

n % 95% CL* n % 95% CL*

Configuration

Annular 40 70 57, 81 31 54 42, 66

Fimbriated 12 21 12, 33 4 7 3, 17

Crescentict 0 0 0, 6 16 28 18, 41

Sleeve 4 7 3, 17 6 11 5, 21

Septated I 2 0, 9 0 0 0, 6

Notch(es) 17 38 25,52 13 29 17,43

External ridge(s)t 47 82 71, 90 8 14 7, 25

Longitudinal intravaginal ridge(s) 31 54 41, 67 30 53 40, 65

Bump(s) 0 0 0,6 3 5 2,14

Tag(s) 3 5 2,14 6 11 5,21

*Confidence limits.

t Indicates significant difference in prevalence between newborn and 1 year with nonoverlapping confidence limits.

(3)

Fig 1. This hymen underwent little morphologic change between birth and 13 months, with persistence of an annular configuration and intravagmal ridges at the 6 and 9 o’clock positions.

Fig 2. An annular configuration with tissue surrounding the hymen 3600 at birth became crescentic by 13 months. Note persistence of the intravagmal ridges laterally and inferiorly (arrows).

at birth. Notches were not observed infenionly

be-tween the 4 and 8 o’clock positions at birth or at 1

year of age.

Longitudinal intravaginal ridges that extended to

the hymenal rim were observed with almost the

same frequency at both ages. However, their location

shifted from commonly being observed superiorly

between the 11 and I o’clock positions in 16% (5/31)

and inferiorly between the 5 and 7 o’clock positions

in 39% (12/31) of the newborns with this finding to

almost entirely laterally in 87% (26/30) of the infants

by I year of age (Fig 2).

External ridges appeared significantly more often

in the newborn period. Eighty-nine percent (40/45)

of newborns who had a ridge at the 6 o’clock position

and 71% (5/7) with one at the 12 o’clock position did

not have this characteristic by their second set of

photographs (Fig 4).

Tags were noted to extend from the rim, an

exter-nal ridge, on an intravaginal ridge. Four girls with a

tag at 1 year had not had this finding previously. In

two, an intravaginal ridge observed infenionly at

birth extended beyond the rim of the hymen and

formed a tag by I year of age. The other two girls at

I year had a tag extending from the rim which

ap-peaned to be a remnant of the redundant tissue at

birth (Fig 4). Tags resolved in two instances and

pen-sisted in the same location in two additional

cases.

Bumps appeared in three i-year-old children, one

of whom was noted to have an external ridge at the

same location at birth and another who previously

had a cleft adjacent to the area of the bump. The third

child had a fimbriated hymen at birth with multiple

clefts which appeared annular at I year with two

(4)

Fig 3. A fimbriated hymen at birth evolved into a crescentic configuration

with an absence of tissue from the 11 to

1 o’clock positions by 12 months of age. Note the vestibular bands at I year of age (arrows).

As the abundant hymenal tissue obscured the

yes-tibuban area at birth, vestibular bands were observed

infrequently in newborns (8%; 95% CL = 7%, 25%)

compared with I-year-obds (95%; 95% CL = 86%,

98%). Additionally, the mean number of bands

ob-served per infant among those who had this finding

was greater at I year than at birth (3.9 vs 2.4). In 75%

(6/8) of newborns and 78% (42/54) of I-yean-olds in

whom the bands were observed, they were located

superior to the hymen on both sides of the urethra.

The second most common location was immediately

adjacent to the peniunethnal bands in 38% (3/8) at

birth and in 82% (44/54) at I year. Vestibulan bands

were not observed laterally in any of the infants at

birth because of the redundancy of the hymenal

tis-sue, but 20% (11 /54) were noted to have lateral

bands at I year. Bands were noted infeniorby in 13%

(1/8)

at birth and 9% (5/54) at I year.

Fig 4. An external ridge at the 6

o’clock position and lateral clefts at the

3 and 9 o’clock positions (arrows) ob-served in this redundant annular hy-men at birth were no longer present by 14 months of age. A tag (arrow) formed superiorly by I year of age as the tissue turned inward.

A lacy vascular pattern on the hymen was

ob-served in all newborns and 1-yean-olds in whom the

vessels could be visualized. No hymenab cysts were

observed at birth or I year in this population.

DISCUSSION

Previous studies of the hymen in nonabused

chil-dren have focused on its appearance in each subject

at a single point in time.3’6 These cross-sectional

studies illustrate the prevalence of notches, bumps,

bands, and ridges at different ages and demonstrate

that hymenal opening sizes and configurations

sig-nificantly vary according to the child’s age. They fail

to show, however, the changes that occur in the

hy-men of a single child over time due to aging and

varying estrogen bevels. This is the first longitudinal

study to document the hymen’s appearance in

non-abused prepubertal girls in the first year of life.

(5)

Consistent with previous publications on the

hy-men with photographic documentation,1’2 findings

reported in this paper are limited to those observable

in photographs. Use of this technique permits

pro-longed scrutiny of the anatomy through examination

of the pictures, yet requires holding the preverbab

child in the examining position for only a limited

amount of time. Photographic documentation also

decreases intenobsenven differences demonstrated to

occur in up to 67% of cases in studies based on

ex-amination alone.7 All children were photographed in

the supine position because of the very young age of

the subjects. The labial traction technique was

se-lected oven the separation method as it has been

demonstrated to be superior for opening the vaginal

introitus.8

The low participation nate (16%) can be attributed

to the fact that the University of Texas Medical

Branch services patients from the entire state of Texas

as webb as from out of state. Follow-up was difficult

because of the barge area to be covered. Several

par-ents expressed an interest in participating but were

unable to travel to the institution. Five additional

infants who had been photographed at 6 weeks of

age under a different protocol were recruited for

photographs at I year of age to augment this study.

Interviewing the child for a history of sexual

as-sault was not an option in this study, as all subjects

were preverbab. Ninety-one percent (52/57) of the

initial photographs were done prior to discharge

from the nursery, thereby essentially eliminating a

question of abuse. Although possible, the chance of

assault prior to the 1-year photographs is low

be-cause of the very young age of the subjects.

Labial agglutination was not observed at birth and

has only rarely been reported as a congenital

find-ing.9 It has been reported in children as an acquired

finding as early as 6 weeks of age1#{176}and occurs more

frequently before 12 months of age than after.2 The

finding of complete on partial agglutination in 13%

(8/62)

of 1-yean-obds who did not have this

chanac-tenistic at birth supports the premise that this is

rarely a congenital trait. The etiology remains

un-clean, however. Although labial adhesions have

pre-viousby been suggested as a marker for sexual

as-saubt,11 the frequency of this finding in studies on

children selected for nonabuse suggests that it

fre-quently occurs in the absence of assault.1’2 Candida

infection of the vulva may predispose the infant to

agglutination, as three of the eight girls with vulvan

fusion had a previous rash. Infection does not

nec-essanily result in fusion, however, since seven

moth-ens whose infants did not have agglutination on

ex-amination also reported a history of a rash.

A redundant hymen, present in all newborns,

per-sisted in 42% (24/57) of I-yean-olds. The tissue of the

majority of infants demonstrated a decrease in

thick-ness and number of folds. Hymens usually appeared

invented in the newborn period because of an

abun-dance of tissue relative to the small vaginal opening.

By I year, it turned inward with the surface that had

previously faced outward now toward the vaginal

canal. This decrease in hymenal tissue in the first

year of life correlates with a decrease in serum

estra-diob bevels. Estrogen levels are high in the neonate,

due to the maternal transmission of hormones in

uteno, but drop dramatically after birth. By

5

days of

life gonadotnopin levels rise in response to estrogen

withdrawal, causing a transient rise in plasma estniol

during

the first few months of life. Gonadotropin

levels then fall until the child nears puberty, although

follicle-stimulating hormone bevels may not be

max-imally suppressed until age 4#{149}12The observation of a

scanty hymen without overlapping folds in the

ma-jonty of infants at I year is consistent with a low

estrogen bevel in most, but not all, girls at this age.

The variation between infants in the amount of

hy-menab tissue present at I year may be due to

mdi-vidual difference in estrogen levels on its effects on

the tissue.

Previous studies document that the crescentic

hy-men is the most common configuration in children

4 years old,2 yet it does not occur in newborns.4 The

longitudinal design of this study shows that this

con-figuration frequently begins as an annular or

fimbni-ated hymen with a superior notch at the 12 o’clock

position that widens to encompass the entire area

between the 11 and I o’clock positions. Ten of the 11

infants who had a superior notch at birth developed

a crescentic hymen by I year of age. Only one child

born with an annular hymen and a notch at the 12

o’clock position had persistence of the

circumfenen-tial configuration at I year. Interestingly, nine girls

with an annular hymen at I year had a superior

notch which was not present at birth. An extended

longitudinal study is necessary to determine whether

these hymens eventually become crescentic.

Lateral notches, which tended to be more

supenfi-cial than those noted superiorly, resolved by 1 year of

age in all but two cases. The rim recedes as the

thick-ness and number of folds in the redundant tissue

decrease. Therefore, the innermost point of the

su-penficial indentations becomes pant of a smooth rim.

Inferior notches, between the 4 and 8 o’clock

posi-tions, were not observed at either age in this study or

in two previous cross-sectional studies on nonabused

girls.1’2 In a comparison of sexually assaulted girls to

normal subjects with no genital complaints and

chil-dnen with other genital complaints, Emans et ab6

ob-served hymenal clefts at the 6 o’clock position in the

abused group only. An inferior notch should,

there-fore, continue to be considered an acquired,

abnon-mab finding.

Longitudinal intnavaginal ridges occurred with

equal frequency but in different locations at birth and

I year. Numerous ridges were observed in the

new-born

vagina which were not classified as

longitudi-nab intravaginab ridges because they did not extend

to on beyond the hymenal rim. As the hymen

be-comes less redundant oven time, these ridges may

reach the rim and meet the criteria of a longitudinal

intravaginal ridge. Additionally, the relative absence

of this finding superiorly and infeniorly at I year may

partially be due to difficulty observing the inner

sun-face of the hymen in these locations in olden children

as compared to the newborn period when the tissue

is inverted.

(6)

Bumps were noted in three I-year-old girls, one

that appeared to be a remnant of a previous external

ridge. The mounds in the other two children

ap-peaned to be remnants of an uneven fimbriated

con-figuration with multiple notches and bumps at birth.

No bumps were observed that could not be traced

back to a similar finding at birth. Although not

ob-served in this study, hymenal mounds on projections

have been reported in nonabused girls in association

with longitudinal intravaginal ridges.1’2 Emans et al6

similarly reported that bumps occurred with equal

frequency in symptomatic, asymptomatic, and

sexu-ably abused girls, but observed that those with an

abuse history more frequently had this finding

infe-nionly between the 3 and 9 o’clock positions. This

report supports that hymenal bumps frequently

oc-cur in the absence of abuse. Further studies with

photographs comparing children with and without a

history of sexual assault are important to determine

whether a relationship exists between trauma and

the presence and location of bumps.

More vestibubar bands were observed at I year

than in the newborn period. This finding is probably

due to the abundant hymenab tissue at birth

obscun-ing the area superior to the hymen rather than an

actual difference in prevalence. Similar to McCann

and coworkers’ study,1 bands were usually bilateral

and most frequently appeared on either side of the

urethral meatus. Lateral and posterior attachments to

the vestibule were also commonly noted.

CONCLUSION

This is the first study to document

photognaphi-cabby the dynamic nature of the hymen’s appearance

in a cohort of infants over a 1-year period. The most

marked change was a decrease in the amount of

hymenab tissue, frequently resulting in a change in

configuration, and the number of bands, notches, or

ridges. Significant increases in the proportion of

in-fants with crescentic configurations and vestibular

bands were noted I year after birth. A significant

decrease in the proportion with external ridges was

also found. No change in the number of infants with

longitudinal intravaginal ridges was observed.

Infe-non clefts between the 4 and 8 o’clock positions were

not seen at birth on I year and should be considered

an abnormal, acquired finding. By better

under-standing the influence that age and changing

estro-gen levels can have on the appearance of the external

genitalia, clinicians will be better equipped to

differ-entiate congenital findings from posttnaumatic

changes.

ACKNOWLEDGMENTS

I thank Anna Somma-Garcia for recruitment and excellent

as-sistance, Dr Steve Barnett for assistance with nurse orientation

and photographs, Dr Astrid Heger for her suggestions, Dr Gregg

Wilkinson for statistical analysis, Dr Constance Wiemann for

ed-iting, and Elizabeth Holt for data entry and manuscript

prepara-tion.

REFERENCES

1. McCann J, Wells R, Simon M, Voris J. Genital findings in prepubertal girls selected for non-abuse: a descriptive study. Pediatrics. 1990;86:428-439

2. Berenson AB, Heger AH, Hayes JM, Bailey RK, Emans SJ. Appearance

of the hymen in prepubertal girls. Pediatrics. 1992;89:387-394

3. Goff CW, Burke KR, Rickenback C, Buebendorf DP. Vaginal opening measurement prepubertal girls. AJDC. 1989;143:1366-1 368

4. Berenson AB, Heger AH, Andrews SA. Appearance of the hymen in

newborns. Pediatrics. 1991;87:458-465

5. Rothman KJ, Boice JD Jr. Epidemiologic Analysis With a Programmable Calculator. Chestnut Hill, MA: Epidemiology Resources Inc; 1982:31-32

6. Emans SJ, Woods ER, Flagg NT, Freeman A. Genital findings in sexually abused, symptomatic and asymptomatic girls. Pediatrics. 1987;79:778-785

7. Chacko MR. Kozinetz CA, Bermudez A. Examination of the hymen in

prepubertal children with suspected sexual abuse: interobserver agree-ment. Adolesc Pediatr Gynecol. 1991 ;4:1 87-193

8. McCann J, Voris J, Simon M, Wells R. Comparison of genital examina-tion techniques in prepubertal girls. Pediatrics. 1990;85:182-187

9. Bowles HE, Childs LS. Synechias of vulva in small children. AJDC.

1953;66:258-263

10. Barysh N. Vulvar fusion in infancy and childhood. N EngI JMed.

1954;250:637-638

11. McCann J, Voris J, Simon M. Labial adhesions and posterior fourchette injuries in childhood sexual abuse. AJDC. 1988;142:659-663

12. Emans SJH, Goldstein DP. Pediatric and Adolescent Gynecology. 3rd ed.

Boston, MA: Little Brown and Company; 1990:95-97

(7)

1993;91;820

Pediatrics

Abbey B. Berenson