Forensic Science International 266 (2016) 80–85

Contents lists available at ScienceDirect

Forensic Science International

journal homepage: www.elsevier.com/locate/forsciint

Adult femicide victims in forensic autopsy in Taiwan: A 10-year

retrospective study
Wen-Li Fong a, Chih-Hsin Pan b, James Chun-I Lee c, Tsui-Ting Lee c, Hsiao-Lin Hwa c,d,e,*
a
Keelung District Prosecutors Office, No. 178, Dungshin Rd., Keelung City, Taiwan ROC
b
Institute of Forensic Medicine, Ministry of Justice, No. 123, Min’an St., Zhonghe Dist., New Taipei City 235, Taiwan, ROC
c
Department and Graduate Institute of Forensic Medicine, College of Medicine, National Taiwan University, No. 1, Sec. 1, Jen Ai Rd., Taipei 100, Taiwan, ROC
d
National Taiwan University, Children and Family Research Center, Leader of Division, No. 1, Sec. 4, Roosevelt Rd., Taipei 106, Taiwan. ROC
e
Department of Obstetrics and Gynecology, National Taiwan University Hospital, No. 7, Chung-Shan South Rd., Taipei 100, Taiwan, ROC

A R T I C L E I N F O A B S T R A C T

Article history: Background: Violence against women and adult femicides are critical medico-legal issues worldwide.
Received 25 April 2016 Intimate partner violence is one of the leading contributory risk factors. This study aimed to describe the
Accepted 10 May 2016 characteristics of femicides in Taiwan.
Available online 17 May 2016
Methods: A retrospective analysis of forensic autopsy records of adult femicide victims in Taiwan during
a 10-year period was carried out. The age, victim-offender relationship, injury patterns and causes of
Keywords: death were analyzed.
Femicide
Results: Among the 220 adult femicide victims recruited, 114 were killed by intimate partners and
Female homicide
Forensic autopsy
106 were killed by non-intimate partner offenders. The average age of victims killed by intimate partners
Intimate partner violence (40.0 y/o) were younger than those killed by non-intimate partner perpetrators (48.6 y/o). The most
Cause of death common site of injuries in the intimate partner group and the non-intimate partner group was the neck
and the upper limbs, respectively. The rates of bruise and intracranial injury of non-intimate partner
group were significantly higher than that of the intimate partner group. The most common causes of
death in both groups were strangulation and sharp force injury. The heart injury was significantly more
frequent in victims offended by intimate partners than by other assailants.
Conclusions: The characteristics of adult femicides, and the patterns of injury in victims killed by
intimate partners and non-intimate partner offenders were different. This data is helpful for corpus
inspection in forensic casework and for strategic planning of femicides prevention.
ß 2016 Elsevier Ireland Ltd. All rights reserved.

1. Introduction
Violence against women is a critical public health problem
worldwide. Femicide (the homicide of women), the extreme type of
violence against women, is one of the leading causes of premature
death among women and an important medico-legal issue [1]. The
incidence and patterns of adult homicide differ from one country to
another. In 2011, there were 4.7 murders per 100,000 in-habitants in
USA, with 22.4% femicide [2]. The homicide rate was 1.561 per
100,000 during 1986–2007 in Taiwan [3]. Femicide accounts for
29.1% of all homicides in 2010 in Taiwan [4]. Despite the magnitude
* Corresponding author at: National Taiwan University, College of Medicine,
Department and Graduate Institute of Forensic Medicine, No. 1, Sec. 1, Jen Ai Rd.,
Taipei 100, Taiwan, ROC. Tel.: +886 2 23123456x65504; fax: +886 2 23218438.
E-mail address: hwahl013@ms10.hinet.net (H.-L. Hwa).
of deaths resulting from violence against women, there are gaps in
data that undermine violence prevention efforts. Many countries
with data lack specificity and the mechanism of femicide, which are
needed to design and monitor prevention efforts. Forensic autopsy
reports provide a specific source of information to document the
characteristics of femicide.
Although the overall risk of homicide for women is lower than
that of men, their risk of being killed by acquaintance, especially
intimate partners, is higher. Contrary to men, the killing of a
woman by a stranger is rare [2]. Intimate partner violence (IPV), a
pivotal legal problem, is the most important risk factor of femicide.
Regardless of gender, anyone may be victims of IPV. However,
females are predominantly the physical IPV victims and predomi-
nantly the IPV homicide victims. Females made up 70% of victims
murdered by an intimate partner (IP) [5]. Previous reports indicate
that women are more likely to be murdered by a current or former
intimate partner than by any other type of perpetrator [5].

http://dx.doi.org/10.1016/j.forsciint.2016.05.008
0379-0738/ß 2016 Elsevier Ireland Ltd. All rights reserved.
 W.-L. Fong et al. / Forensic Science International 266 (2016) 80–85 81

Different characteristics between IPV and general violent
crimes have been described based on studies of nonfatal violence
and murder [6,7]. Given that the motivation and methods of IPV
and non-IPV are different, the pattern of injuries of IP and non-IP
homicides may be varied. Analysis based on the nature of victim-
offender relationships, especially intimate partner (IP) versus
non-intimate partner (non-IP), allows investigators to explore
different forms of femicides of these two groups. Besides, the
types of femicide are closely related to culture, and weapon usage
as well as injury patterns may vary in different countries. In Asia,
in-depth analysis on femicide is limited. However, these
researches are needed for authorities to design and monitor
prevention efforts.
This study was designed to examine the forensic autopsy
records of adult femicide cases in Taiwan. We aimed to analyze
demographic characteristics, victim-offender relationship, meth-
ods of injury, patterns of injuries, and causes of death (COD) among
adult femicide victims.
2. Materials and methods
This retrospective study was approved by the Institutional
Review Board and was carried out at an institute that executes
about 85.5% of medico-legal autopsies in Taiwan. The records of all
available adult female homicides (aged 18 and above) brought to
this institute for medico-legal autopsy from 2001 to 2010 and
taken as solved cases with known identity of the suspects were
examined. Data were collected from forensic autopsy records,
including results of scene investigation. Detailed information
regarding the perpetrators and circumstances of the crime was
reported by the police under prosecutors’ supervision. The medico-
legal autopsy was performed by forensic pathologists, with
histological examination as a routine procedure. Toxicological
analyses were conducted on victims routinely, unless the analysis
was regarded as being unfeasible due to severe decomposition of
the corpus, blood transfusion or other special reasons. DNA tests
were performed for individual identification, when necessary. The
data collected in our work included victim identity, gender, age,
place of the event, relationship between the victim and the
murderer(s), gender of the killer(s), type of weapon used, the
anatomical region(s) injured, type(s) and number of injuries, the
COD, and the manner of death. Strangulation was defined as all
types of asphyxia generated by manual strangulation, and ligature
strangulation [8]. Cases were classified into IP femicide and non-IP
femicide. IP femicide is defined as the killing of a woman by an IP
(husband, ex-husband, boyfriend, or ex-boyfriend). Non-IP femi-
cide is defined as the killing of a woman by non-IP offender(s).
Descriptive analyses were carried out using the Statistical
Product and Service Solutions (SPSS) software (Version 22.0.0.0,
SPSS Inc., Chicago, USA). Descriptive statistics (means and
proportions) were generated. Significance was determined using
the Pearson’s x2 test and independent sample t test, as appropriate.
P values less than 0.05 were considered statistically significant.
3. Result
During a 10-year period from January 2001 to December 2010,
16,160 autopsies were performed at this institute, and homicides
accounted for about 14.0% of the forensic autopsy cases, with 30.7%
femicides. In all, 220 adult female cases (aged 18 and above, mean
age 44.1  17.0) with homicide as the manner of death, were
recruited in this study. Table 1 presents the demographic data and
vulnerability condition of female homicides classified by IP and non-
IP attackers (other attackers). There were 114 (51.8%) cases of IP
femicides and 106 cases of non-IP femicides. The average age of
victims of IP femicide group was 40.0 (13.8) year-old, whereas the
average age of victims of other perpetrators group was 48.6 (18.8)
year-old, with a significant difference (p < 0.001). The age distribu-
tion of these two groups is illustrated in Fig. 1. Most of victims killed
by IPs were 20–49 y/o. Among victims murdered by other attackers,
the age distribution was relatively even at age 20–69. Among victims
attacked by IPs, the most common perpetrator was the boyfriend
(53.5%), followed by the husband (29.8%), ex-boyfriend (10.5%), and
ex-husband (6.1%). Among the victims murdered by non-IPs, the
offenders included relative(s) (34.9%), friend(s) (29.4%), other
acquaintance(s) (29.4%), and stranger(s) (6.6%).
Supplementary Table S1 demonstrates the available demo-
graphic data of the assailants. All IP offenders were male, whereas
14.7% female attackers were recorded in the non-IP group
(p < 0.001). There were three victims murdered by two perpe-
trators. Criminal records were documented in 20.2% of IP offenders
and 35.8% of non-IP perpetrators (p = 0.009). About 12.8% of non-IP
offenders had mental illness, which was significantly higher than
that of IP murderers (4.4%) (p = 0.024).
Table 2 demonstrates the investigation and toxicology results of
these adult femicide victims. In both groups, knives were the most
commonly used weapons, followed by bodily force (Table 2).
Firearms were used in three cases killed by IPs. Histology (data not
shown) and toxicology were performed in 100.0% and 78.6% of
cases, respectively. Alcohol consumption was detected in 20.2%

Table 1
Demographic data and vulnerability of adult femicide victims killed by intimate partners and non-intimate partner offenders.

IP % (no.) Non-IP % (no.) p-Value All % (no.)

N = 114 N = 106 N = 220

Age 40.0  13.8 48.6  18.8 <0.001* 44.1  17.0

Marriage 0.815
Never married 22.8 (26) 18.9 (20) 20.9 (46)
Married 41.2 (47) 39.6 (42) 40.5 (89)
Other 21.9 (25) 26.4 (28) 24.1 (53)
Unknown 14 (16) 15.1 (16) 14.5 (32)
Occupation 0.111
Student 2.6 (3) 3.8 (4) 3.2 (7)
Unemployed/housewife 44.7 (51) 51.9(55) 48.2 (106)
Employed 36.8 (42) 38.7 (41) 37.7 (83)
Unknown 15.8 (18) 5.7 (6) 10.9 (24)
Vulnerability
Aged (365 y/o) 5.3 (6) 24.5 (26) <0.001* 14.5 (32)
Mental illness 1.8 (2) 2.8 (3) 0.636 2.3 (5)
Disabled 0.9 (1) 2.8 (3) 0.304 1.8 (4)
Alcohol/drug abuse 4.4 (5) 3.8 (4) 0.819 4.1 (9)

IP: intimate partner; Non-IP: non-intimate partner.

*
p-value < 0.05.
82 W.-L. Fong et al. / Forensic Science International 266 (2016) 80–85

Table 2 18 cases (15.8%) attacked by IP, and in one case (0.9%) offended by
Investigation and toxicology results of adult femicides.
non-IP perpetrators (p < 0.001).
IP % (no.) Non-IP % (no.) p-value All % (no.) In IP groups, the most common injuries were bruise (52.6%),
N = 114 N = 106 N = 220 followed by cut (49.1%), stab (45.6%), intra-thoracic injury (35.1%),
Location 0.923 and fracture (31.6%). (Table 3) In non-IP groups, the most common
Victim’s home 58.8 (67) 54.7 (58) 56.8 (125) injuries were bruise (67.9%), followed by cut (44.3%), stab (40.6%),
Offender’s home 9.6 (11) 9.4 (10) 9.5 (21) fracture (35.8%), and intracranial injuries (31.1%). Among all kinds
Public places 7.0 (8) 8.5 (9) 7.7 (17)
of injuries, bruise and intra-cranial injury were significantly more
Other 21.1 (24) 25.5 (27) 23.2 (51)
Methods 0.270 frequent in victims attacked by non-IP perpetrators than by IPs.
Bodily force 24.6 (28) 20.8 (22) 22.7 (50) Table 4 illustrates the distribution of injury of body regions of
Poison 6.1 (7) 1.9 (2) 4.1 (9) these cases. Among victims offended by IPs, the neck (58.8%), face
Weapon(s) 76.3 (87) 78.3 (83) 77.3 (170)
(56.1%), and upper limbs (53.5%) were the most frequently
Knife 50 (57) 52.8 (56) 51.4 (113)
Gun 2.6 (3) 0.0 (0) 1.4 (3)
involved body regions. For victims attacked by other perpetrators,
Rope 4.4 (5) 8.5 (9) 6.4 (14) upper limbs (66.0%), the face (65.7%), and cranium (62.3%) were
Stick 6.1 (7) 5.7 (6) 5.9 (13) the most frequently wounded body regions. The rate of cranial
Other objects 16.7 (19) 13.2 (14) 15.0 (33) trauma in victims offended by non-IP perpetrators was signifi-
Sexual assault 0.9 (1) 9.4 (10) 0.004* 5.0 (11)
cantly higher than that by IPs (p = 0.002).
Murder suicide 15.8 (18) 0.9 (1) <0.001* 8.6 (19)
Toxicology exam 82.5 (94) 74.5 (79) 78.6 (173) Table 5 presents the COD of the victims. Among victims
Positive 31.9 (30/94) 25.3 (20/79) 0.340 28.9 (50/173) perpetrated by IPs, the most common COD was strangulation/
Type smothering (27.2%), followed by sharp force lung trauma (22.8%)
Sedatives 5.3 (5/94) 6.3 (5/79) 0.745 5.8 (10/173)
and sharp force neck vessels injury (15.8%). The commonest COD
Illicit drugs 4.3 (4/94) 6.3 (5/79) 0.514 5.2 (9/173)
Alcohol 20.2 (19/94) 11.4 (9/79) 0.117 16.2 (28/173)
was strangulation/smothering (22.6%), followed by sharp force
CO 2.1 (2/94) 0.0 (0/79) 0.171 1.2 (2/173) lung injury (17.9%) and blunt force head injury (17.9%) in victims
Resticide 0.0 (0/94) 1.3 (1/79) 0.274 0.6 (1/173) killed by non-IP offenders. There was significantly higher rate of
IP: intimate partner; Non-IP: non-intimate partner. sharp force heart injury as COD in victims killed by IPs in this study.
*
p-value < 0.05. There were more blunt force head injury, sharp force head injury,
[(Fig._1)TD$IG] and sharp force abdominal injury as COD in victims murdered by
non-IPs. In sexual homicide cases, strangulation/suffocation
(45.5%, 5/11) were the most common COD. Among murder-suicide
victims, the most frequent COD was sharp force trauma (47.4%, 9/
19), followed by strangulation (42.1%, 8/19). The most frequent
fatal injury regions were neck in both groups (Supplementary
Table S2). There was significantly more fatal thoracic injury in IP
group, and significantly more fatal cranial injury in non-IP group

4. Discussion

Information from the present study contributes to our

Fig. 1. Age distribution of femicide victims killed by intimate partners and non-
intimate partner perpetrators.
and 11.4% of victims attacked by IPs and non-IPs, without
significant difference. Sexual homicide was documented in one
victim (0.9%) offended by IPs and in ten victims (9.4%) attacked by
other perpetrators (p = 0.004). Murder-suicides were confirmed in
understanding of the characteristics of adult femicides in Taiwan.
Approximately half of female victims were killed by IPs in this
study. It is correlated with a report by the WHO [9]. Previous
studies indicate that IPV offenders, compared to those who
committed non-intimate partner violence, had less lifetime
antisocial behavior, less other nonviolent offending, less illicit
drug abuse, and more self-constraint [6,10]. It is similar to our
findings with less criminal records of offenders in IP group than
those in non-IP group. These two groups displayed different
motives for committing murder in a previous report. Those who
murdered intimate partners were more likely to be motivated to

Table 3
Injury types of adult femicide victims.

Types IP % (no.) Non-IP % (no.) p-value All % (no.)

N = 114 N = 106 N = 220

Bruise 52.6 (60) 67.9 (72) 0.021* 60.0 (132)

Hematoma 7.0 (8) 13.2 (14) 0.126 10.0 (22)
Abrasion 28.9 (33) 24.5 (26) 0.460 26.8 (59)
Cut 49.1 (56) 44.3 (47) 0.477 46.8 (103)
Stab 45.6 (52) 40.6 (43) 0.450 43.2 (95)
Fracture 31.6 (36) 35.8 (38) 0.503 33.6 (74)
Burn 3.5 (4) 0.9 (1) 0.202 3.3 (5)
Laceration 13.2 (15) 22.6 (24) 0.066 17.7 (39)
Intracranial injury 19.3 (22) 31.1 (33) 0.043* 25.0 (55)
Intra-thoracic organs or vessels injury 35.1 (40) 27.4 (29) 0.217 31.4 (69)
Intra-abdominal organs or vessels injury 7.0 (8) 11.3 (12) 0.267 9.1 (20)

IP: intimate partner; Non-IP: non-intimate partner.

*
p-value < 0.05.
 W.-L. Fong et al. / Forensic Science International 266 (2016) 80–85 83

Table 4 commonly used weapon. Our results were correlated with a

Distribution of injuries on body regions of adult femicide victims.
previous report with stabbing, beating, and strangulation as
Body region IP % (no.) Non-IP % (no.) p-value All % (no.) common personal forms of violence [16]. However, 76.3% of IP
N = 114 N = 106 N = 220 offenders used weapons in this study. It is different from previous
Face 56.1 (64) 65.1 (69) 0.175 60.5 (133) reports suggesting that beating to death was most common in IP
Upper 30.7 (35) 36.8 (39) 0.339 33.6 (74) femicides [17,18]. Methods of murder and weapons usage in
Middle 41.2 (47) 51.9 (55) 0.113 46.4 (102) homicides vary in different countries [17,19]. Firearms were the
Lower 32.5 (37) 32.1 (34) 0.952 32.3 (71)
most common weapons used in femicides in USA [2]. In our study,
Cranium 41.2 (47) 62.3 (66) 0.002* 51.4 (113)
Neck 58.8 (67) 47.2 (50) 0.085 53.2 (117) only three victims were killed with firearms. The strict firearm
Chest 50.9 (58) 49.1 (52) 0.787 50.0 (110) legislation in Taiwan was the reason. There is a very low rate of gun
Abdomen 21.9 (25) 28.3 (30) 0.275 25.0 (55) ownership in Taiwan, so firearms were uncommon weapons in this
Back 25.4 (29) 22.6 (24) 0.628 24.1 (53)
study. The prevalence of firearm ownership was reported to be
Buttocks 1.8 (2) 3.8 (4) 0.358 2.7 (6)
Upper limbs 53.5 (61) 66.0 (70) 0.059 59.5 (131)
positively related to the incidence of firearm homicide [20].
Lower limbs 31.6 (36) 37.7 (40) 0.337 34.5 (76) Bruises were the most frequent external injuries documented in
Ano-genitalia 1.8 (2) 6.6 (7) 0.070 4.1 (9) this study; followed by cut and stab. Although knives were
IP: intimate partner; Non-IP: non-intimate partner. commonly used in cases of both groups, blunt force trauma that
*
p-value < 0.05. includes bruise, abrasion, hematoma, and laceration were also
common in victims of both groups [8]. The fracture rate of IP
femicide victims noted in this study (31.6%) was higher than that in
kill by emotional concerns (hatred and rage) versus practical previous reports of IPV survivors [21,22]. It is compatible to the
concerns (money, related to other crimes) [7]. Because fatal severity of physical violence leading to death in our cases. Among
violence against women offended by IPs are dissimilar in their women ever injured through IPV, the fracture rate was reported to
causes from other perpetrators, prevention strategies of femicide be 21.9% in a population-based survey of New Zealand [22]. That
should be evaluated and reformulated account for victim-offender report was based on interviewing of population-based cluster
relationship. sampled female survivors without ascertaining injuries by medical
Femicides represents a significant source of premature records. The possibility of recall bias and an underestimate of
mortality among women, as well as the impacts on family frequency of occurrence may also be the reasons of a lower fracture
survivors including children consequently [11]. The age distribu- rate in that report.
tion of adult femicide victims in this study includes a relatively The anatomical distribution of trauma showed a predominance
high incidence of women aged 30–49, particularly among women of neck, facial, and upper limb injuries among victims attacked by
murdered by IPs. It is compatible with the age distributions in IPs in this study. On the other hand, upper limb, face, and cranial
previous reports of femicide, with a peak incidence in the women injuries were commonest among victims killed by non-IP
aged 30–49 [12,13]. Among the victims murdered by IPs in this perpetrators. The association between the distribution of the
study, more than half were aged 30–49, with a peak at women aged wounds and the victim-perpetrator relationship has been dis-
40–49. The distribution was different from the mandatory cussed [18,23]. The face may be targeted in homicides with
reporting data of Taiwan, with a peak at women aged 30–39 emotional meaning [24]. Trojan and Krull concluded that injuries
[14]. The majority of the cases in the mandatory reports were to the head and face were more likely among IP offenders [18]. A
survivors. The age distribution of the IPV survivors and IP femicide meta-analysis demonstrated that head, neck or facial injuries are
victims could be different. Most of IP femicide victims experienced significant markers for IPV [25]. Facial injuries, particular
physical violence for a long time before being killed. The likelihood maxillofacial, or middle facial injuries have been indicated to be
that IP femicides declined sharply after the age of 50 is similar to associated with IPV and IPV homicide [18,23,26]. However, Last
the previous report of IP victims in ambulatory care settings [15]. and Fritzon concluded that facial wounding occurred nearly
The majority of victims in both groups were attacked with sharp equally among strangers, family member and IP homicides
objects, followed by bodily force, with knives being the most [19]. The link between wounded regions and victim-offender
relationship was examined in this study. The facial injury rate was
similarly high among our IP and non-IP femicide victims, with
Table 5
Causes of death of adult femicide victims.
middle face as the most commonly involved region. In this study,
the rate of facial injury of victims offended by IPs was similar to a
Cause of death IP % (no.) Non-IP % (no.) p-value All % (no.) previous report [26]. The rate of middle face injury of IP group was
N = 114 N = 106 N = 220
lower than that of non-IP group in this study, without significance.
Sharp object injury 52.6 (60) 53.8 (57) 0.824 53.2 (117) Facial wounding may be part of a violent attack in victims
Head 0.9 (1) 7.5 (8) 0.013* 4.1 (9)
perpetrated by acquaintance, relative, or intimate partner who are
Neck vessels 15.8 (18) 15.1 (16) 0.887 15.5 (34)
Lung 22.8 (26) 17.9 (19) 0.370 20.5 (45) furious. Besides, facial injuries were more common among those
Heart 10.5 (12) 2.8 (3) 0.024* 6.8 (15) who did resist compared to those who did not [27]. The battery
Abdomen 1.8 (2) 7.5 (8) 0.039* 4.5 (10) patterns may vary in different cultures. Punching style attack is
Other 0.9 (1) 2.8 (3) 0.279 1.8 (4) common in some western countries, while relatively rare in our
Strangulation/smothering 27.2 (31) 22.6 (24) 0.436 25 (55)
Blunt force injury 13.2 (15) 20.8 (22) 0.132 16.8 (37)
culture [28]. In this study, the connection between facial wounding
Head 7.9 (9) 17.9 (19) 0.026* 12.7 (28) and IP femicide was not clear. The neck injuries were not
Other 5.3 (6) 2.8 (3) 0.363 4.1 (9) significantly more common in IP femicide victims than that of
Poison 2.6 (3) 1.9 (2) 0.711 2.3 (5) non-IP femicide victims, and significantly higher rate of cranial
Firearm 2.6 (3) 0.0 (0) 0.093 1.4 (3)
injury was noted in victims murdered by non-IPs. The utility of the
Pushing from height 0.0 (0) 0.9 (1) 0.300 0.5 (1)
Motor vehicle injury 0.9 (1) 0.0 (0) 0.334 0.5 (1) head, neck, and facial injury patterns to predict the relationship to
Drowning and submersion 0.9 (1) 0.0 (0) 0.334 0.5 (1) their offender in investigation may be limited in our country.
IP: intimate partner; Non-IP: non-intimate partner.
The neck was the most common fatal injury region in both
*
p-value < 0.05. group, because strangulation and sharp force neck injury were
common in both groups. The significantly more frequent thoracic
84 W.-L. Fong et al. / Forensic Science International 266 (2016) 80–85
fatal injury in IP group and the significantly more frequent cranial
fatal injury in non-IP group suggested the different patterns of
injury distribution in these two groups. Previous studies indicated
that the head was the most common injury site in blunt force
homicides and the neck and chest were the most common sites of
sharp force homicides [29–31]. It is similar to our findings. The
high mortality rate of trauma involving the blood vessels of the
neck and lungs or heart of the chest may be the reasons [29].
In this study, strangulation/smothering and sharp force injury
were the most frequent COD in both groups. The physical
disadvantage of female victims led to the failure of resistance,
because most of the offenders were male and manual strangulation
was one of the common methods [32]. Sharp force heart injury was
a significantly more common COD among IP femicide victims,
while sharp force abdominal injury was a significantly more
common COD among non-IP femicide victims in this study. Severe
heart injury and abdominal organs/vessels injuries were docu-
mented in 28% and 12% of sharp force homicide victims,
respectively [33]. Heart and abdominal organs injuries were
common COD of sharp instrument fatalities [33,34]. However,
thoracic and abdominal injuries were relatively uncommon in
nonfatal cases, and were nonspecific for IPV survivors presented to
emergency departments [25]. The injury patterns of fatal and non-
fatal attacks by IPs or non-IPs were different. Besides, intentional
heart injury to commit murder by an IP may have specific meaning
for the offender. To our knowledge, there was no previous
literature reviewing sharp force heart injuries as COD by IPs.
Further research can be performed to evaluate the difference of
COD between IP and non-IP femicide.
Sexual homicides comprised 5.0% of femicide in this study. This
percentage is similar to those of previous reports from the USA
[35,36]. In our study, sexual homicide was determined to have
occurred in 0.9% and 9.4% of victims attacked by IPs and non-IPs,
respectively. However, in a previous report, about 44% of sex-
related homicide victims were killed by their IPs [37]. The evidence
of sexual assault is easily overlooked, and the incidence of sexual
homicide in both groups could be underestimated, especially in the
IP group [35–37]. The inclusion of experts in sexual assault
examination responding to femicide and the utilization of
technology to improve the detection of ano-genital trauma as
well as the collection of trace evidence can enhance the
identification of sex-related femicide [37]. Strangulation was the
leading cause of death in sexual homicide cases, followed by sharp
force injury in this study. It is consistent with previous reports with
strangulation, blunt force trauma, and stabbing as the most
frequent COD for sex-related homicide [36,38]. The close physical
contact with the victim facilitates manual strangulation performed
by the assailant, when the offender wanted to kill to prevent the
victim from reporting the sexual assault or when the offender used
excessive force to keep the victim from screaming [36–38].
Murder-suicide cases were noted in 15.8% of victims attacked
by IPs and 0.9% of victims offended by non-IPs in this study. It is
similar to previous reports of IP homicide in our country and in
South Africa [39,40]. However, the rate is lower than that reported
by Bossarte et al., with 30.6% of male IP homicide perpetrators
committed suicide afterwards [41]. Most of the murder-suicide
victims were killed by sharp instruments and asphyxiation in this
study. The methods of murder and suicide vary in different
countries [42,43]. Similar to our result, the rate of murder-suicide
was relatively lower under tough gun laws, and the primary
methods of murder were strangulation and sharp force trauma in
this kind of cases in England and Wales [42].
Violence against women is a public health issue, more effective
policies to prevent lethal violence should be designed and
implemented. The forensic examiners may have a significant role
in the reorganization and in the combating of violence against
women including IPV. Given that the method used and the injury
patterns may be different between IP and non-IP homicides, the
intervention efforts to prevent homicide in general population may
be ineffective for IP murders. The individual and private nature of
the IP homicide suggests multi-pronged approach addressing
feelings and risk factors within intimate relationships may be
helpful for IP homicides [7]. From our study, education to women
suffered from IPV with prevention of neck holding and prevention
of accessing of knives by their IPs may be effective. Base on the
wound patterns of femicides, the forensic physicians can help
primary health care providers understand the cause and effect of
injury, and offering appropriate medical management (for exam-
ple: comprehensive examination of neck, skeletal image study,
examination for intra-thoracic and intra-abdominal injury) for
survivors at emergency department. Complete physical examina-
tion and forensic evidence collection for possible sexual assault of
femicide victims are suggested. This data can also help police
officers in dealing with fatal and non-fatal violence against women
and voluntary supporters’ organizations in developing prevention
strategy forums.
In conclusion, the present study demonstrated the differences
of age distribution, characteristics of cases, injury types, anatomi-
cal distribution of injuries, and COD between adult femicides
committed by IPs and non-IPs in Taiwan. The results of this study
will increase health providers’ and forensic examiners’ awareness
of femicides. The characteristics of injuries may help to improve
clinical management of survivors as well as the death investigation
of women, and to launch effective preventive strategies.
Acknowledgements
This study was partially supported by National Taiwan
University Children and Family Research Center. The authors
thank the Department of Forensic Pathology of Institute of Forensic
Medicine, Ministry of Justice, R.O.C. (Taipei, Taiwan) for informa-
tion supplement.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.forsciint.2016.05.
008.
References
[1] J.C. Campbell, D. Webster, J. Koziol-McLain, C. Block, D. Campbell, M.A. Curry, F.
Gary, N. Glass, J. McFarlane, C. Sachs, P. Sharps, Y. Ulrich, S.A. Wilt, J. Manganello,
X. Xu, J. Schollenberger, V. Frye, K. Laughon, Risk factors for femicide in abusive
relationships: results from a multisite case control study, Am. J. Public Health 93
(2003) 1089–1097.
[2] Federal Bureau of Investigation, Uniform Crime Reports, Department of Justice,
Washington, DC, 2011, https://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/
2011/crime-in-the-u.s.-2011/violent-crime/murder (accessed 21.03.16).
[3] S.W. Yu, L. Pai, S.H. Tsai, C.C. Lai, Y.H. Huang, W.C. Chien, Homicide mortality
trends by sex, age and cause in Taiwan, 1986–2007, Chung Shan Med. J. 21 (2010)
47–59.
[4] Department of Statistics, Ministry of Health and Welfare, ROC, http://www.
mohw.gov.tw (accessed 21.02.16).
[5] S. Catalano, E. Smith, H. Snyder, M. Rand, Female Victims of Violence, U.S.
Department of Justice Office of Justice Programs, Bureau of Justice Statistics, 2009.
[6] T.E. Moffitt, R.F. Krueger, A. Caspi, J. Fagan, Partner abuse and general crime: how
are they the same? How are they different?, Criminology 38 (2000) 199–232.
[7] K.A. Thomas, M.E. Dichter, J. Matejkowski, Intimate versus non-intimate partner
murder: a comparison of offender and situational characteristics, Homicide Stud.
15 (2011) 291–311.
[8] J. Fischer, W.J. Kleemann, H.D. Tröger, Types of trauma in cases of homicide,
Forensic Sci. Int. 68 (1994) 161–167.
[9] H. Stöckl, K. Devries, A. Rotstein, N. Abrahams, J. Campbell, C. Watts, C.G. Moreno,
The global prevalence of intimate partner homicide: a systematic review, Lancet
382 (2013) 859–865.
 W.-L. Fong et al. / Forensic Science International 266 (2016) 80–85
[10] D.J. Boyle, D. O’Leary, A. Rosenbaum, C. Hassett-Walker, Differentiating between
generally and partner-only violent subgroups: lifetime antisocial behavior, family
of origin violence, and impulsivity, J. Fam. Violence 23 (2008) 47–55.
[11] V. Frye, S. Wilt, Femicide and social disorganization, Violence Against Women 7
(2001) 335.
[12] Statistics Canada, CANSIM, Homicide Survey, Canadian Centre for Justice Statis-
tics, 2015, http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/
legal10a-eng.htm (accessed 22.02.16).
[13] M. Lee-Gorman, S. MacNeill, D. Rizet, S.D. McDermott, Homicide/suspicious death
statistics for cases submitted to the forensic science laboratory in the Republic of
Ireland for 2004–2008, Med. Sci. Law 51 (2011) 146–150.
[14] Department of Statistics, the Ministry of the Interior, ROC, http://www.moi.gov.
tw/stat/week.aspx (accessed 21.02.16).
[15] E. Petridou, A. Browne, E. Lichter, X. Dedoukou, D. Alexe, N. Dessypris, What
distinguishes unintentional injuries from injuries due to intimate partner vio-
lence: a study in Greek ambulatory care settings, Inj. Prev. 8 (2002) 197–201.
[16] R.A. Silverman, S.K. Mukherjee, Intimate homicide: an analysis of violent social
relationships, Behav. Sci. Law 5 (1987) 37–47.
[17] H. Johnson, T. Hotton, Losing control: homicide risk in estranged and intact
intimate relationships, Homicide Stud. 7 (2003) 58–84.
[18] C. Trojan, A.C. Krull, Variations in wounding by relationship intimacy in homicide
cases, J. Interpers. Violence 27 (2012) 2869.
[19] S.K. Last, K. Fritzon, Investigating the nature of expressiveness in stranger,
acquaintance and intrafamilial homicides, J. Investig. Psych. Offender Profil. 2
(2005) 179–193.
[20] UNODC (United Nations Office on Drugs and Crime), Global Study on Homicide
2013 Trends, 2014, http://www.unodc.org/documents/gsh/pdfs/2014_GLOBAL_
HOMICIDE_ BOOK_web.pdf (accessed 10.12.15).
[21] M. Bhandari, S. Dosanjh, P. Tornetta 3rd, D. Matthews, Musculoskeletal manifes-
tations of physical abuse after intimate partner violence, J. Trauma 61 (2006)
1473–1479.
[22] J.L. Fanslow, E.M. Robinson, Physical injuries resulting from intimate partner
violence and disclosure to healthcare providers: results from a New Zealand
population-based study, Inj. Prev. 17 (2011) 37–42.
[23] B.T. Le, E.J. Dierks, B.A. Ueeck, L.D. Homer, B.F. Potter, Maxillofacial injuries
associated with domestic violence, J. Oral Maxillofac. Surgery 59 (2001) 1277–
1283.
[24] J.R. Meloy, Empirical basis and forensic application of affective and predatory
violence, Aust. N. Z. J. Psychiatry 40 (2006) 539–547.
[25] V. Wu, H. Huff, M. Bhandari, Pattern of physical injury associated with intimate
partner violence in women presenting to the emergency department: a system-
atic review and meta-analysis, Trauma Violence Abuse 11 (2010) 71–82.
85
[26] N. Saddki, A.A. Suhaimi, R. Daud, Maxillofacial injuries associated with intimate
partner violence in women, BMC Public Health 10 (2010) 268.
[27] K. Fritzon, J. Ridgway, Near-death experience: the role of victim reaction in
attempted homicide, J. Interpers. Violence 16 (2001) 679–696.
[28] J.C. Campbell, P. Sharps, N.E. Glass, Risk assessment for intimate partner violence,
in: G.F. Pinard, L. Pagani (Eds.), Clinical Assessment of Dangerousness: Empirical
Contributions, Cambridge University Press, New York, 2001, pp. 136–157.
[29] K.I. Au, S.L. Beh, Injury patterns of sharp instrument homicides in Hong Kong,
Forensic Sci. Int. 204 (2011) 201–204.
[30] V.N. Ambade, H.V. Godbole, Comparison of wound patterns in homicide by sharp
and blunt force, Forensic Sci. Int. 156 (2006) 166–170.
[31] S. Rogde, H.P. Hougen, K. Poulsen, Homicide by sharp force in two Scandinavian
capitals, Forensic Sci. Int. 109 (2000) 135–145.
[32] S. Rogde, H.P. Hougen, K. Poulsen, Asphyxial homicide in two Scandinavian
capitals, Am. J. Forensic Med. Pathol. 22 (2001) 128–133.
[33] C.J. Kemal, T. Patterson, D.K. Molina, Deaths due to sharp force injuries in Bexar
County, Texas, with respect to manner of death, Am. J. Forensic Med. Pathol. 34
(2013) 253–259.
[34] J.R. Gill, C. Catanese, Sharp injury fatalities in New York City, J. Forensic Sci. 47
(2002) 554–557.
[35] J.R. Meloy, The nature and dynamics of sexual homicide: an integrative review,
Aggress. Violent Behav. 5 (2000) 1–22.
[36] I.T. Van Patten, P.Q. Delhauer, Sexual homicide: a spatial analysis of 25 years of
deaths in Los Angeles, J. Forensic Sci. 52 (2007) 1129–1141.
[37] T. Henry, Characteristics of sex-related homicides in Alaska, J. Forensic Nurs. 6
(2010) 57–65.
[38] N. Abrahams, L.J. Martin, R. Jewkes, S. Mathews, L. Vetten, C. Lombard, The
epidemiology and the pathology of suspected rape homicide in South Africa,
Forensic Sci. Int. 178 (2008) 132–138.
[39] S. Mathews, N. Abrahams, R. Jewkes, L.J. Martin, C. Lombard, L. Vetten, Intimate
femicide-suicide in South Africa: a cross-sectional study, Bull. World Health Org.
86 (2008) 552–558.
[40] P.L. Wang, Intimate partner femicide, Formosa J. Ment. Health 25 (2012) 231–266.
[41] R.M. Bossarte, T.R. Simon, L. Barker, Characteristics of homicide followed by
suicide incidents in multiple states, 2003–2004, Inj. Prev. 12 (2006) 33–38.
[42] A. Travis, L. Johnson, C. Milroy, Homicide-suicide (dyadic death), homicide and
firearms use in England and Wales, Am. J. Forensic Med. Pathol. 28 (2007) 314–
318.
[43] J. Logan, H.A. Hill, M.L. Black, A.E. Crosby, D.L. Karch, J.D. Barnes, K.M. Lubell,
Characteristics of perpetrators in homicide-followed-by-suicide incidents: na-
tional violent death reporting system – 17 US States, 2003–2005, Am. J. Epide-
miol. 168 (2008) 1056–1064.