Results

We conducted a systematic review of the scientific literature on the epidemiology and criminology of femicide for comparative analysis. Following the PRISMA-guided selection process (Fig. 1), sixteen retrospective studies were included in the final analysis. The main characteristics of the included studies—namely country of origin, study period, study design, sample size, availability of autopsy data, and psychiatric or criminological variables—are summarized in Table 1. The studies were published between 2004 and 2024 and covered a wide geographical distribution, including North America, Europe, South America, the Middle East, Asia, and Africa. Although all included articles were published within this time frame, several studies analyzed long retrospective series with observation periods extending back to the mid-20th century. Study designs were heterogeneous and included population-based registries, forensic autopsy series, and retrospective analyses of medico-legal or judicial records [37, 38].

Author (year) | Country | Study period | Study design | IPF / NIPF | Sample size | Autopsy data | Psychiatric /criminological data | Main findings | Key limitations
Belfrage & Rying (2004) | Sweden | 1990–1999 | National retrospective | IPF | 175 | Partial | Yes | IPF associated with prior violence and jealousy | Registry-based data, limited psych detail
Henry (2010) | Alaska (USA) | 1990–2008 | Retrospective | IPF / sex-related | 64 | Partial | Yes | Firearms predominant | Small sample
Dobash & Dobash (2011) | UK | 1990–2008 | Qualitative / case review | IPF | 122 | No | Yes | Control and jealousy central | Narrative design
Kivivuori & Lehti (2012) | Finland | 1990–2009 | Population-based | IPF / other homicide | 189 | No | Limited | IPF differs from other homicides in relational dynamics | Lack of autopsy detail
Liem et al. (2013) | Finland, NL, Sweden | 2003–2006 | Multinational registry | IPF | 577 | No | Limited | Firearms less prevalent in Nordic countries | Heterogeneous datasets
Fong et al. (2016) | Taiwan | 2003–2013 | Autopsy-based | IPF | 153 | Yes | Limited | Sharp-force most frequent | Single-country
Meneghel et al. (2017) | Brazil | 2003–2007 | Population-based | IPF | 521 | No | Limited | High alcohol involvement | Underreporting
Toprak & Ersoy (2017) | Turkey | 2000–2010 | Registry-based | IPF | 129 | No | Limited | Strangulation and beating common | Media-based bias
Vatnar et al. (2017) | Norway | 1990–2012 | Retrospective forensic | IPF | 177 | Partial | Yes | Separation as major risk factor | Retrospective design
Salameh et al. (2018) | Jordan | 2005–2015 | Autopsy-based | IPF / NIPF | 102 | Yes | No | Sharp-force and strangulation prevalent | Limited psych data
Cunha & Gonçalves (2019) | Portugal | 2000–2014 | Retrospective | IPF | 89 | No | Yes | Substance abuse and prior convictions | Small sample
Zara et al. (2019) | Italy | 1993–2013 | Retrospective forensic | IPF / NIPF | 113 | Yes | Yes | Distinct criminological profiles | Regional data
Macucha & Taunde (2020) | Mozambique | 2016–2018 | Autopsy-based | IPF / NIPF | 81 | Yes | No | Non-partner killings prevalent | Short study period
Bologna study (2022) | Italy | 1950–2020 | Autopsy-based | IPF | 276 | Yes | Limited | Long-term trend analysis | Historical heterogeneity
Franchetti et al. (2024) | Germany / Italy | 2000–2020 | Autopsy-based | IPF | 131 | Yes | Partial | Injury patterns and weapon use | Forensic focus only
Lewis et al. (2024) | USA | 2003–2020 | National surveillance | IPF | > 2000 | No | Yes | High prevalence of firearms and prior DV | Registry-based

Across the included studies focusing on intimate partner femicide (IPF), the perpetrator was most frequently identified as the victim’s current or former intimate partner, including husbands, cohabiting partners, or long-term romantic partners. In high-income countries population-based and registry studies, intimate partners accounted for approximately 75–80% of femicide perpetrators, with limited variability across countries [8, 39, 40]. In contrast, a different distribution was reported in specific regional contexts, such as Mozambique, where non-intimate perpetrators accounted for up to 76% of cases [41].

When the setting of the homicide was reported, the event most commonly occurred in a domestic environment, including the victim’s home or a shared residence. Population-based studies from the United States described a broader distribution of homicide locations, including public spaces and vehicles, although domestic settings remained predominant [42].

Marked geographical variability was observed in the methods of killing. In studies conducted in the United States and other high income countries, firearms represented the most frequently reported cause of death, accounting for approximately 40–50% of intimate partner femicides, followed by sharp-force injuries such as stabbing or cutting [37, 42]. In contrast, studies from low–middle income countries reported higher frequencies of strangulation, blunt-force trauma, and poisoning. In Mozambique and Turkey, strangulation accounted for approximately 25–30% of reported cases [41, 43]. Asian forensic autopsy series identified sharp-force injuries as the leading cause of death, followed by strangulation and physical beating [44, 45].

Several autopsy-based studies reported the presence of multiple injury mechanisms in the same victim, including combinations of strangulation, blunt-force trauma, and sharp-force injuries, reflecting complex assault dynamics [46, 47]. Defensive injuries, such as abrasions, contusions, and fractures of the upper limbs, were frequently described when detailed forensic examinations were available.

Information on psychiatric and criminological characteristics of perpetrators was variably reported across the included studies. Severe mental disorders, including psychotic disorders, were documented in approximately 5–10% of perpetrators in studies providing psychiatric data [34, 48]. A broader category of psychiatric symptomatology at the time of the offense, including depressive or psychotic features, was reported in up to 20% of cases in some forensic and registry-based studies [8]. Substance abuse emerged as a recurrent finding across multiple geographical areas. Chronic alcohol or drug abuse was reported in approximately 15–20% of perpetrators overall, with alcohol being the most frequently involved substance, followed by illicit drugs such as cocaine, methamphetamines, marijuana, and heroin [49]. Higher proportions of perpetrators with positive toxicological findings or documented substance abuse histories were reported in specific national studies, including South Africa, Portugal, England, and Brazil, with rates ranging between 30% and 40% [48, 49].

Several studies examined temporal and relational dynamics preceding the homicide. When data were available, the period following relationship separation or termination emerged as a particularly high-risk interval. In cases involving former partners, a substantial proportion of intimate partner femicides occurred within the first year after separation, with some studies reporting rates exceeding 50% [40, 50].

Due to substantial heterogeneity in study design, populations, outcome definitions, and data sources, no formal quantitative synthesis or meta-analysis was performed. All numerical values presented represent descriptive ranges derived from specific subsets of studies, including population-based registries, forensic autopsy series, or national reports, rather than pooled global estimates.