Sexual Abuse of Children The Detection of Semen Tina
on
Skin
Gabby, MD, MPH; Marilyn A. Winkleby, PhD; W. Thomas Boyce, MD; George F. Sensabaugh, D Grim
Deborah L. Fisher, MPH; Allison Lancaster;
Objective.\p=m-\The detection of semen on the skin of children who present within 72 hours of an episode of sexual assault is critical to medical, forensic, and legal personnel. The Wood's Lamp, a UV light that causes semen to fluoresce, and four forensic laboratory techniques were compared to determine their sensitivity and decline in sensitivity over time. \s=b\
Design.\p=m-\A descriptive study. Participants. \p=m-\Eleven adult female volunteers.
Measurements/Main Results.\p=m-\Semen was placed on the skin of the volunteers. Samples of the dried semen were assessed during a 28-hour period with the Wood's Lamp, microscopy, the acid phosphatase assay, and two assays for the prostatic protein p30 (counterimmunoelectrophoresis and incidence studies reveal that approximately National form of of all children will experience It estimated that of child ual abuse 1%
some
sex¬
45% in a given year.1 is sexual assault cases involve acts of molestation with gen¬ ital contact without penile penetration.1 Ejaculation may occur in these cases, making the presence of semen on a child's skin diagnostically important. Unfortunately, sen¬ sitivities of the routine screening and diagnostic tests for detecting semen on skin are unknown. These tests are used in acute cases of sexual abuse, cases seen within 72 hours of an alleged assault, and are not considered to be of value after that time. Because children often present to a medical examiner days after an episode of sexual abuse, it is important to know how increasing time affects the results of these tests.2 It is critical that clinicians and forensic scientists, understand the limitations of the diagnostic and screen¬ ing tests used during the examination of the sexually abused child. This study was conducted to evaluate the sensitivity and specificity of the Wood's Lamp (WL) and the sensitivities of four diagnostic techniques used to identify semen on a victim's skin. The WL, a screening tool used within 72 hours of an assault, creates a UV illumination that detects semen flu¬ orescence.3 Detection of semen fluorescence is especial-
Accepted for publication January 3, 1992. From the Division of Behavioral and Developmental Pediatrics, University of California, San Francisco (Drs Gabby, Winkleby, and
Boyce) and the Forensic Science Group, School of Public Health, University of California, Berkeley (Mss Fisher and Lancaster and Dr
Sensabaugh).
Presented at the Maternal and Child Health Bureau\p=n-\sponsored Annual Meeting of Behavioral Pediatric Training Programs, San Francisco, Calif, April 12, 1991. Reprints not available.
enzyme-linked immunosorbent assay). The intensity of the
Wood's Lamp fluorescence of semen diminished dramatically by 28 hours; in contrast, the fluorescence of urine persisted up to 80 hours. Over time, the p30\p=m-\enzyme-linked immunosorbent assay technique was more sensitive than microscopy, the acid phosphatase assay, and p30\x=req-\ counterimmunoelectrophoresis in detecting semen on skin. Conclusions. \p=m-\TheWood's Lamp is not a sensitive screening tool and should be used with caution. To improve the detection of sexual abuse in children, we recommend that the p30\p=m-\enzyme-linkedimmunosorbent assay be used because of its potential as a more sensitive assay than those in current clinical use.
(AJDC. 1992;146:700-703)
ly helpful in cases in which a young child's ability to give a history of sexual assault may be limited. Wood's Lamp fluorescence is not specific to semen since many other
substances that may be found on skin have fluorescent its widespread availability and sim¬ WL the plicity, technique is not standardized, and there have been no studies to date regarding its speci¬ ficity or sensitivity over time. Typically, when a fluo¬ rescent area is visualized during the examination of an abused child, a swab or scraping of the area is ob¬ tained.4 The swab is further analyzed for semen by forensic analyses. Three forensic techniques commonly used in the anal¬ yses of semen evidence were concurrently examined in this study: (1) sperm detection by microscopy, (2) the as¬ say for acid phosphatase (ACP), and (3) the detection of the prostate-specific protein, p30, by counterimmunoelectrophoresis (CIE). A fourth technique, the detection of p30 by an enzyme-linked immunosorbent assay (ELISA), was also examined. These tests vary in their sensitivity and specificity. Detection of spermatozoa by microscopy is considered diagnostically specific for sexual contact.5 Acid phosphatase is less specific since it is present in the vaginal fluid and urine of women, albeit at much lower levels than in semen.6-7 It is not known if ACP is present in the vaginal fluid or urine of prepubescent girls; it is present at very low levels in the urine of prepubescent boys.8 Negative ACP test results have no diagnostic value since they do not necessarily indicate the absence of se¬ men.9 Prostate-specific protein, p30 (also called prostatespecific antigen), is found in seminal fluid and urine of men, and under normal circumstances is absent from all other body fluids in males and from all tissues and body fluids in females.10-11 The p30-ELISA test is approximately 100 times more sensitive than the ACP or p30-CIE tests. Although these semen markers have been studied in
properties. Despite
Downloaded From: http://archpedi.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 05/18/2015
postcoital vaginal samples and their specificity has been outlined, little is known about the decay in their sensitiv¬ ities over time when samples are obtained from skin. The following questions were addressed by this study: (1) How does the pattern, color, and intensity of WL flu¬ orescence of semen change over time? (2) Do nonsemen substances, likely to be present on the skin of children re¬ ferred for sexual abuse examinations, cause false-positive WL readings (thereby compromising specificity)? (3) How does the sensitivity of microscopy, ACP, and p30 tech¬ niques, analyzed from semen samples taken from the ar¬ eas of WL fluorescence, change over time? (4) Can foren¬ sic tests detect semen in areas that are WL negative? Although adult subjects participated in this study, our intent was to address these questions as they specifically relate to the examination of child sexual assault victims.
SUBJECTS AND METHODS
Preliminary Testing
Preliminary testing
was conducted to provide general esti¬ of time and body location on the WL fluo¬ rescence of semen. Since observations of semen on skin during an 80-hour period demonstrated that semen fluorescence be¬ came faint after the first 28 hours, this period was chosen for the study. There was no remarkable difference in semen fluores¬ cence and semen analysis when semen placed on the inner thigh was compared with semen placed on the ventral forearm; there¬ fore, the forearm was chosen for the site of study. Preliminary testing was also conducted to identify substances that might be present in child sexual assault cases and have the potential of causing false-positive WL readings. These included infant formula, milk, cola, moisturizing lotions, petroleum jelly, a common lubricating jelly (K-Y Jelly, Johnson & Johnson, Skillman, NJ), a contraceptive foam and cream, and urine. The flu¬ orescence of most of these substances differed from semen flu¬ orescence in color, texture, or pattern. Urine and formula exhibited a color and pattern of fluorescence similar to semen. Moreover, urine fluorescence persisted through the preliminary observation period of 80 hours. Because urine has the potential of being present in the permeai area of young children and the fluorescence of urine could be confused with that of semen, urine was included in the study to examine its potential for gen¬ erating false-positive WL readings.
mates of the effects
Ultraviolet Products Ine, San Gabriel, Calif) over the forearms, and the fluorescence of semen and urine was recorded as either positive, faint, or negative. The pattern of fluorescence was re¬ corded as even or irregular, and the color was noted as yellowgreen or white. Each subject had a WL reading performed at 15 minutes and 4, 7, 11, 24, and 28 hours after semen and urine application. At each interval, specimens were taken for forensic analyses by swabbing one of the semen areas with a cotton-tipped applica¬ tor moistened with sterile water. These swabs were air dried, frozen at 20°C, and transported to the laboratory. Control swabs, taken from areas of the skin without semen or urine applications, were collected from 10 (91%) of the 11 sub¬ jects. Because the specificities of the forensic assays were known, only three (27%) of the 11 subjects had swabs of urine col¬ lected. These samples were similarly air dried and frozen for later analyses.
Laboratory Analyses All forensic examiners were blind to the timing of the swab sample collection. The cotton portions of the swabs were extracted in isotonic Hepes buffer and centrifuged to separate the supernatant from the cellular debris; both fractions were fro¬ zen for later analyses. Microscopy was performed using the cellular debris from the above procedure. A so-called Christmas tree stain12 was used that stains sperm heads red and tail structures green. Results of microscopy were recorded as positive or negative depending on the presence or absence of spermatozoa. The ACP activity was assayed with a standard quantitative assay using p-nitrophenyl phosphate as a substrate.9 The cutoff point for significance of ACP values was defined at 0.1 IU/mL since all skin control val¬ ues were below this threshold. The CIE assay for p30 allows semiquantitation of p30 levels and is sensitive to approximately 500 ng/mL.13 The p30-CIE re¬ sults were recorded as positive or negative. The p30-ELISA as¬ say allows detection of p30 at concentrations as low as 2 ng/mL.14 All skin control values registered less than the 2 ng/mL detection limit for the assay configuration used (Epichrome Immunometric Assay for prostatic antigen, Cetus Corp, Emeryville, Calif); therefore, this was used as our cutoff point below which test re¬ sults were considered negative. (Although the assay kit used in this study is no longer commercially available, comparable assays for prostatic antigen p30 are available from Yang Labora¬ tories, Seattle, Wash, and Hybritech, La Jolla, Calif.)
Subjects and Sample Collection
RESULTS
Eleven women were recruited from two universities in north¬ ern California during February and March 1990, and informed consent for participation in the study was obtained. Each subject obtained a semen sample from a male partner on the evening prior to or on the morning of the study. Samples collected in the evening were promptly refrigerated after collection. (During preliminary testing, the fluorescent and forensic properties of refrigerated semen were found to be no different than those of fresh semen.) On the morning of their study participation, sub¬ jects were asked to avoid the use of lotions or creams on their arms and to collect a sample of their own urine. During the 28hour study period, subjects were asked to avoid bathing and to wear clothing that protected their arms during the day to min¬ imize friction against their skin.
(91%) of the 11 subjects had WL evaluations and swabs of dried semen taken at baseline and at the five specified intervals to 28 hours after semen application; the remaining subject missed one WL reading and swab col¬ lection 4 hours after application. One subject also missed one WL reading but was instructed to collect a swab sam¬ ple per study protocol. In all, 65 samples were analyzed. The visual appearance of semen and urine spots on the subjects' skin was assessed during the 28-hour study pe¬ riod. Semen spots initially appeared shiny or colloid-like, then turned flaky (72% of subjects had a flaky appearance 4 hours after semen application), and finally cleared over time. After 24 hours, the skin with semen appeared clear in nine subjects (82%). In contrast, skin with urine appli¬ cations appeared clear throughout the study period.
Test
Applications and Observations
Liquid semen in 50-µ aliquots was placed on eight demar¬ cated, ventral forearm locations of each volunteer. Fiftymicroliter samples of urine were also applied to two areas on the
forearms. One semen area and one urine area served as sites for the measurements of fluorescence decay, as described below. The other sites were reserved for sample collection for forensic analyses. During the 28-hour test period the same examiner (T.G.) held a WL (Mineralight UVL-56 Blak Ray Ultraviolet Lamp,
Ten
WL
Findings
The color of semen fluorescence was indistinguishable from the color of urine fluorescence throughout all peri¬ ods. The pattern of semen fluorescence appeared irregu¬ lar in texture, while the pattern of urine fluorescence was consistently homogeneous over time. As shown in Fig 1,
Downloaded From: http://archpedi.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 05/18/2015
on
Skin
Gabby, MD, MPH; Marilyn A. Winkleby, PhD; W. Thomas Boyce, MD; George F. Sensabaugh, D Grim
Deborah L. Fisher, MPH; Allison Lancaster;
Objective.\p=m-\The detection of semen on the skin of children who present within 72 hours of an episode of sexual assault is critical to medical, forensic, and legal personnel. The Wood's Lamp, a UV light that causes semen to fluoresce, and four forensic laboratory techniques were compared to determine their sensitivity and decline in sensitivity over time. \s=b\
Design.\p=m-\A descriptive study. Participants. \p=m-\Eleven adult female volunteers.
Measurements/Main Results.\p=m-\Semen was placed on the skin of the volunteers. Samples of the dried semen were assessed during a 28-hour period with the Wood's Lamp, microscopy, the acid phosphatase assay, and two assays for the prostatic protein p30 (counterimmunoelectrophoresis and incidence studies reveal that approximately National form of of all children will experience It estimated that of child ual abuse 1%
some
sex¬
45% in a given year.1 is sexual assault cases involve acts of molestation with gen¬ ital contact without penile penetration.1 Ejaculation may occur in these cases, making the presence of semen on a child's skin diagnostically important. Unfortunately, sen¬ sitivities of the routine screening and diagnostic tests for detecting semen on skin are unknown. These tests are used in acute cases of sexual abuse, cases seen within 72 hours of an alleged assault, and are not considered to be of value after that time. Because children often present to a medical examiner days after an episode of sexual abuse, it is important to know how increasing time affects the results of these tests.2 It is critical that clinicians and forensic scientists, understand the limitations of the diagnostic and screen¬ ing tests used during the examination of the sexually abused child. This study was conducted to evaluate the sensitivity and specificity of the Wood's Lamp (WL) and the sensitivities of four diagnostic techniques used to identify semen on a victim's skin. The WL, a screening tool used within 72 hours of an assault, creates a UV illumination that detects semen flu¬ orescence.3 Detection of semen fluorescence is especial-
Accepted for publication January 3, 1992. From the Division of Behavioral and Developmental Pediatrics, University of California, San Francisco (Drs Gabby, Winkleby, and
Boyce) and the Forensic Science Group, School of Public Health, University of California, Berkeley (Mss Fisher and Lancaster and Dr
Sensabaugh).
Presented at the Maternal and Child Health Bureau\p=n-\sponsored Annual Meeting of Behavioral Pediatric Training Programs, San Francisco, Calif, April 12, 1991. Reprints not available.
enzyme-linked immunosorbent assay). The intensity of the
Wood's Lamp fluorescence of semen diminished dramatically by 28 hours; in contrast, the fluorescence of urine persisted up to 80 hours. Over time, the p30\p=m-\enzyme-linked immunosorbent assay technique was more sensitive than microscopy, the acid phosphatase assay, and p30\x=req-\ counterimmunoelectrophoresis in detecting semen on skin. Conclusions. \p=m-\TheWood's Lamp is not a sensitive screening tool and should be used with caution. To improve the detection of sexual abuse in children, we recommend that the p30\p=m-\enzyme-linkedimmunosorbent assay be used because of its potential as a more sensitive assay than those in current clinical use.
(AJDC. 1992;146:700-703)
ly helpful in cases in which a young child's ability to give a history of sexual assault may be limited. Wood's Lamp fluorescence is not specific to semen since many other
substances that may be found on skin have fluorescent its widespread availability and sim¬ WL the plicity, technique is not standardized, and there have been no studies to date regarding its speci¬ ficity or sensitivity over time. Typically, when a fluo¬ rescent area is visualized during the examination of an abused child, a swab or scraping of the area is ob¬ tained.4 The swab is further analyzed for semen by forensic analyses. Three forensic techniques commonly used in the anal¬ yses of semen evidence were concurrently examined in this study: (1) sperm detection by microscopy, (2) the as¬ say for acid phosphatase (ACP), and (3) the detection of the prostate-specific protein, p30, by counterimmunoelectrophoresis (CIE). A fourth technique, the detection of p30 by an enzyme-linked immunosorbent assay (ELISA), was also examined. These tests vary in their sensitivity and specificity. Detection of spermatozoa by microscopy is considered diagnostically specific for sexual contact.5 Acid phosphatase is less specific since it is present in the vaginal fluid and urine of women, albeit at much lower levels than in semen.6-7 It is not known if ACP is present in the vaginal fluid or urine of prepubescent girls; it is present at very low levels in the urine of prepubescent boys.8 Negative ACP test results have no diagnostic value since they do not necessarily indicate the absence of se¬ men.9 Prostate-specific protein, p30 (also called prostatespecific antigen), is found in seminal fluid and urine of men, and under normal circumstances is absent from all other body fluids in males and from all tissues and body fluids in females.10-11 The p30-ELISA test is approximately 100 times more sensitive than the ACP or p30-CIE tests. Although these semen markers have been studied in
properties. Despite
Downloaded From: http://archpedi.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 05/18/2015
postcoital vaginal samples and their specificity has been outlined, little is known about the decay in their sensitiv¬ ities over time when samples are obtained from skin. The following questions were addressed by this study: (1) How does the pattern, color, and intensity of WL flu¬ orescence of semen change over time? (2) Do nonsemen substances, likely to be present on the skin of children re¬ ferred for sexual abuse examinations, cause false-positive WL readings (thereby compromising specificity)? (3) How does the sensitivity of microscopy, ACP, and p30 tech¬ niques, analyzed from semen samples taken from the ar¬ eas of WL fluorescence, change over time? (4) Can foren¬ sic tests detect semen in areas that are WL negative? Although adult subjects participated in this study, our intent was to address these questions as they specifically relate to the examination of child sexual assault victims.
SUBJECTS AND METHODS
Preliminary Testing
Preliminary testing
was conducted to provide general esti¬ of time and body location on the WL fluo¬ rescence of semen. Since observations of semen on skin during an 80-hour period demonstrated that semen fluorescence be¬ came faint after the first 28 hours, this period was chosen for the study. There was no remarkable difference in semen fluores¬ cence and semen analysis when semen placed on the inner thigh was compared with semen placed on the ventral forearm; there¬ fore, the forearm was chosen for the site of study. Preliminary testing was also conducted to identify substances that might be present in child sexual assault cases and have the potential of causing false-positive WL readings. These included infant formula, milk, cola, moisturizing lotions, petroleum jelly, a common lubricating jelly (K-Y Jelly, Johnson & Johnson, Skillman, NJ), a contraceptive foam and cream, and urine. The flu¬ orescence of most of these substances differed from semen flu¬ orescence in color, texture, or pattern. Urine and formula exhibited a color and pattern of fluorescence similar to semen. Moreover, urine fluorescence persisted through the preliminary observation period of 80 hours. Because urine has the potential of being present in the permeai area of young children and the fluorescence of urine could be confused with that of semen, urine was included in the study to examine its potential for gen¬ erating false-positive WL readings.
mates of the effects
Ultraviolet Products Ine, San Gabriel, Calif) over the forearms, and the fluorescence of semen and urine was recorded as either positive, faint, or negative. The pattern of fluorescence was re¬ corded as even or irregular, and the color was noted as yellowgreen or white. Each subject had a WL reading performed at 15 minutes and 4, 7, 11, 24, and 28 hours after semen and urine application. At each interval, specimens were taken for forensic analyses by swabbing one of the semen areas with a cotton-tipped applica¬ tor moistened with sterile water. These swabs were air dried, frozen at 20°C, and transported to the laboratory. Control swabs, taken from areas of the skin without semen or urine applications, were collected from 10 (91%) of the 11 sub¬ jects. Because the specificities of the forensic assays were known, only three (27%) of the 11 subjects had swabs of urine col¬ lected. These samples were similarly air dried and frozen for later analyses.
Laboratory Analyses All forensic examiners were blind to the timing of the swab sample collection. The cotton portions of the swabs were extracted in isotonic Hepes buffer and centrifuged to separate the supernatant from the cellular debris; both fractions were fro¬ zen for later analyses. Microscopy was performed using the cellular debris from the above procedure. A so-called Christmas tree stain12 was used that stains sperm heads red and tail structures green. Results of microscopy were recorded as positive or negative depending on the presence or absence of spermatozoa. The ACP activity was assayed with a standard quantitative assay using p-nitrophenyl phosphate as a substrate.9 The cutoff point for significance of ACP values was defined at 0.1 IU/mL since all skin control val¬ ues were below this threshold. The CIE assay for p30 allows semiquantitation of p30 levels and is sensitive to approximately 500 ng/mL.13 The p30-CIE re¬ sults were recorded as positive or negative. The p30-ELISA as¬ say allows detection of p30 at concentrations as low as 2 ng/mL.14 All skin control values registered less than the 2 ng/mL detection limit for the assay configuration used (Epichrome Immunometric Assay for prostatic antigen, Cetus Corp, Emeryville, Calif); therefore, this was used as our cutoff point below which test re¬ sults were considered negative. (Although the assay kit used in this study is no longer commercially available, comparable assays for prostatic antigen p30 are available from Yang Labora¬ tories, Seattle, Wash, and Hybritech, La Jolla, Calif.)
Subjects and Sample Collection
RESULTS
Eleven women were recruited from two universities in north¬ ern California during February and March 1990, and informed consent for participation in the study was obtained. Each subject obtained a semen sample from a male partner on the evening prior to or on the morning of the study. Samples collected in the evening were promptly refrigerated after collection. (During preliminary testing, the fluorescent and forensic properties of refrigerated semen were found to be no different than those of fresh semen.) On the morning of their study participation, sub¬ jects were asked to avoid the use of lotions or creams on their arms and to collect a sample of their own urine. During the 28hour study period, subjects were asked to avoid bathing and to wear clothing that protected their arms during the day to min¬ imize friction against their skin.
(91%) of the 11 subjects had WL evaluations and swabs of dried semen taken at baseline and at the five specified intervals to 28 hours after semen application; the remaining subject missed one WL reading and swab col¬ lection 4 hours after application. One subject also missed one WL reading but was instructed to collect a swab sam¬ ple per study protocol. In all, 65 samples were analyzed. The visual appearance of semen and urine spots on the subjects' skin was assessed during the 28-hour study pe¬ riod. Semen spots initially appeared shiny or colloid-like, then turned flaky (72% of subjects had a flaky appearance 4 hours after semen application), and finally cleared over time. After 24 hours, the skin with semen appeared clear in nine subjects (82%). In contrast, skin with urine appli¬ cations appeared clear throughout the study period.
Test
Applications and Observations
Liquid semen in 50-µ aliquots was placed on eight demar¬ cated, ventral forearm locations of each volunteer. Fiftymicroliter samples of urine were also applied to two areas on the
forearms. One semen area and one urine area served as sites for the measurements of fluorescence decay, as described below. The other sites were reserved for sample collection for forensic analyses. During the 28-hour test period the same examiner (T.G.) held a WL (Mineralight UVL-56 Blak Ray Ultraviolet Lamp,
Ten
WL
Findings
The color of semen fluorescence was indistinguishable from the color of urine fluorescence throughout all peri¬ ods. The pattern of semen fluorescence appeared irregu¬ lar in texture, while the pattern of urine fluorescence was consistently homogeneous over time. As shown in Fig 1,
Downloaded From: http://archpedi.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 05/18/2015
