Stepping Forward, with Caution: The Inter-American Court’s Approach to Transfemicide in Leonela Zelaya

Just a few weeks ago, the Inter-American Court of Human Rights delivered its long-awaited judgment in Leonela Zelaya and Other v Honduras, concerning Leonela Zelaya’s transfemicide.

Leonela Zelaya was born in Cortés, on Honduras’ northern coast. Estranged from her family and expelled from her home, she moved to Tegucigalpa in her twenties, where she engaged in sex work alongside Thalía Rodríguez, another trans woman with whom she developed a sister-like bond. For almost a decade, the two shared a single room in conditions of extreme poverty. At the time of her killing in 2004, Leonela was illiterate and living with HIV and tuberculosis. She was only 34 years old. In 2022, Thalía herself was also murdered.

Two features of what was at stake made this case unusually charged. First, since Vicky Hernández v Honduras (2021), the Court had not had a further opportunity to reaffirm that the Convention of Belém do Pará applies to violence against trans women. Would the Court’s new composition preserve a standard that had once so fiercely divided the tribunal? Second, the Inter-American Commission on Human Rights and Leonela’s representatives pursued an ambitious litigation strategy, urging the Court to attribute direct State responsibility for the transfemicide itself, much as it had done in Vicky Hernández. The difficulty, however, was that the factual configuration in Leonela Zelaya was not entirely analogous, and the evidentiary record was far less robust.

Although the judgment prompted six separate opinions, I argue that Leonela Zelaya ultimately shows that the legacy of Vicky Hernández remains intact—and even more unified than in 2021. In this post, I develop two central claims. First, the Court takes four important steps that further strengthen its jurisprudence on violence affecting LGBTIQ+ persons: (i) the reaffirmation of the Convention of Belém do Pará’s protective scope for trans women; (ii) a broader conception of reparations grounded in affective bonds; (iii) the recognition of gender identity even after death; and (iv) a strengthened intersectional approach to understanding structural violence against trans persons. Second, the judgment offers a cautionary signal, clarifying the evidentiary shortcomings that will prevent the Court from attributing State responsibility for violations of the right to life of women, including trans women. Before turning to these points, I summarise the facts.

The Facts

Between 1998 and 2004, Leonela was detained by police officers on at least three occasions, reportedly for “drunkenness”, “vagrancy”, and “causing a public disturbance”. On 6 September 2004, four days after her last detention, Leonela and Thalía went out to work together, but only Thalía returned home. The next morning, Leonela’s body was found in the street, her chest exposed.

Initial reports indicated that she had been killed with a bladed weapon during the night. Yet she was improperly classified as an “unknown” male, despite the authorities’ knowledge of her identity. The press covered the crime through stereotypical tropes, referring to her as “homosexual” and disclosing her name assigned at birth. No blood relatives attended her funeral or burial. Only the community of trans women around her did—including her sister, Thalía.

The criminal investigation, opened in 2004, yielded little progress. Worse still, the case file remained lost until 2018. Against this backdrop and without exhausting domestic remedies, Leonela’s representatives lodged the case before the Inter-American Commission in 2012. In 2019, Honduran authorities reported that they had identified the perpetrator, but that he had died in 2007. They nevertheless reframed the crime as an attempted robbery rather than as a transfemicide. The case ultimately reached the Court in 2024.

Four Steps Forward

First, the judgment reaffirms that the Convention of Belém do Pará protects trans women (para. 65). This point, perhaps the one that generated the greatest concern ahead of the ruling, ensures that Inter-American jurisprudence remains at the forefront of international protection against violence and discrimination targeting trans women.

The way in which the Court framed the operative part of the judgment (a 4–3 vote), however, does not make entirely clear which of the seven judges embraced this view (para. 195). In fact, the Court reached a 6–1 majority in favour of applying the Convention of Belém do Pará to violence against trans women—improving upon the 5–2 majority on this very question in Vicky Hernández. This time, the sole dissent came from Judge Borea, consistent with his scepticism towards evolutive interpretation and non-textual methodological approaches, even in cases concerning cisgender women.

For those interested in the debate on the extension of the Convention of Belém do Pará to trans women, Judge Moreno’s concurring opinion is noteworthy. Moreno engages directly with then Judges Odio-Benito and Vio-Grossi—the two who dissented in Vicky Hernández. Moreno’s opinion takes up the challenge of responding explicitly to these objections and concludes that their arguments were insufficient to exclude trans women from the treaty’s protective scope. On the contrary, he argues, the evolutive vocation of international human rights law renders such an exclusion untenable (paras. 10–12). In what may become a line of enduring significance, he writes that “those who deny the protection afforded by [the Convention of Belém do Pará] to trans women bear the burden of convincing us that depriving them of such protection would better safeguard their rights” (fn. 16).

Second, the judgment expands the notion of beneficiaries of reparations for breaches of the right to personal integrity, recognising harm suffered by those who do not share a blood relationship with the victim (para. 136). The Court held that Thalía Rodríguez was herself a victim of a violation of personal integrity:

“as a result of the suffering caused by the unjustified delay and the lack of heightened due diligence in the investigation into what happened to Leonela Zelaya, who was her closest bond” (para. 138).

This is a significant development, as it allows the Court to incorporate notions of affective ties in future cases where, as often occurs for LGBTIQ+ persons, chosen family differs from biological family. The Court suggests that victim status cannot be confined to traditional understandings of the family but instead turns on the presence of a genuine and effective bond of affection and care. This move subtly recalibrates the Court’s reparations framework towards a more socially grounded understanding of harm.

At the same time, the reasoning offered remains somewhat narrow, seemingly limited to cases involving trans women who are victims of lethal violence and lack blood relatives. The Court held that:

 “where no family ties exist […], in cases of the violent death of trans women, members of the deceased’s support networks may be recognised as victims of a violation of their mental or moral integrity, provided that a close bond with the victim and an impact on their rights—such as through efforts to obtain justice—can be established” (para. 136).

This raises important questions about the judgment’s broader reach, including whether the same principle would apply beyond cases where no blood relatives survive. Here, the Court also stopped short of engaging more directly with the notion of “social” or chosen family as such.

Judge Pérez-Manrique’s separate opinion adds two important contributions to the majority opinion. First, he argues that Leonela and Thalía’s relationship was protected under Article 17 of the American Convention (protection of the family), since “not considering Thalía a victim would disregard the particular forms that affective and relational life may take”, given that “sometimes bonds of friendship forge relationships of greater cohesion, unity, and support, complemented by cohabitation or accompaniment” (para. 121). Second, he notes that failing to recognise this bond also denied Thalía’s right to the truth and hence her ability to pursue judicial avenues to clarify what happened to her sister (paras. 126–127).

Third, the judgment clearly establishes that States have a duty to respect trans persons’ chosen names in post-mortem investigations (para. 116). Honduran authorities opened a criminal investigation without referring to Leonela as such, despite knowing her gender identity (para. 42; fns. 68, 168). In countries without gender identity laws facilitating recognition of social name and gender, authorities often rely on birth-assigned markers even after death. Here, the Court consolidates an important standard: even in the painful process of clarifying such crimes, States must respect the identity of those who have suffered these extreme forms of violence, a principle it had already foreshadowed in Vicky Hernández (paras. 122–124).

The Court also provides a basis for its practice of naming cases involving trans victims not merely by surname but by their full chosen names—hence Leonela Zelaya v Honduras, rather than simply Zelaya v Honduras (as also in Azul Rojas-Marín v Peru and Vicky Hernández v Honduras). This stands in marked contrast to the European Court of Human Rights’ practice of referring to trans applicants only by initials, and more recently even using birth-assigned gender markers, despite their expressed wishes to the contrary, as in T.H. v the Czech Republic (2025, para. 4).

Fourth, the judgment advances the Court’s use of intersectional tools to understand the discrimination and violence suffered by trans persons (paras. 80–81). In Leonela Zelaya, the Court notes that her gender identity as a trans woman, her role as a sex worker, her extreme poverty, her illiteracy, and her health status gave rise to a combination of factors that not only operated simultaneously but reinforced one another, deepening her exclusion, discrimination, and exposure to multiple forms of violence.

Yet, as in other cases, this intersectional analysis does not fully translate into measures of non-repetition—a limitation recently highlighted by Mónica Arango-Olaya, Óscar Parra-Vera, and Antonio Franco-Franco. In Leonela Zelaya, the Court merely reiterates to Honduras the measures already ordered in Vicky Hernández, noting that they remain pending compliance (paras. 166–169). This was a missed opportunity. While the transfemicides of Leonela and Vicky share important similarities, they also involve distinct layers that might have justified a more tailored approach. Perhaps the Court, aware of the measures of non-repetition still pending in Vicky Hernández, chose not to further complicate the supervisory landscape for Honduras.

A Cautionary Signal

Yet Leonela Zelaya also lends itself to a more sceptical reading—one that begins with the Court’s refusal to attribute direct State responsibility for the transfemicide itself (para. 96), as it had done in Vicky Hernández. Instead, the Court’s most significant findings of State responsibility arose from (i) the failure to investigate Leonela’s death with heightened due diligence, and (ii) the three occasions on which she had been detained by the police under Articles 99 and 100 of the Law on Police and Social Coexistence, provisions contained in a section tellingly entitled Vagancia y Pandillerismo (Vagrancy and Gang Activity).

As regards the latter, the Court held that the vague and open-ended language of Articles 99 and 100—referring to “vagrants”, “itinerant prostitutes”, and “pimps”—breached Articles 7 (personal liberty) and 9 (principle of legality) of the American Convention (paras. 67–78). On that basis, the police actions against Leonela were deemed unlawful and arbitrary (para. 86). The Court offered a careful analysis of these three episodes, concluding that they gave rise to a presumption of discrimination which Honduras had failed to rebut, considering the indicia of negative stereotyping based on Leonela’s gender identity and expression (para. 85).

The Court grounded this conclusion in a four-part structure: (i) Leonela’s heightened vulnerability, understood through an intersectional lens; (ii) the broader context of discrimination and violence against trans women at the time of the detentions (drawing notably on the context found in Vicky Hernández); (iii) the absence of objective elements justifying Leonela’s three detentions; and (iv) the State’s own admission at the hearing that Articles 99 and 100 of the Law on Police and Social Coexistence were ambiguous and therefore susceptible to abusive application by the authorities (para. 86).

It is here, however, that the Court’s true divisions became visible. By a narrow 4–3 majority, it held that Honduras should amend Articles 101 (“drunken persons causing scandal”) and 142 (“offences against modesty, good customs, and public morality”) of the same law—provisions that were similarly vague, but which had not in fact been invoked in Leonela’s detentions (para. 170). This issue did not arise with respect to Articles 99 and 100, which had already been struck down by Honduras’ own Supreme Court in 2015 precisely on grounds of ambiguity. Judges López, Pérez-Goldberg, and Borea expressed disagreement in separate opinions, arguing that the Court erred in ordering legislative reform of provisions that, however clearly incompatible with the American Convention, had not been applied in the case. What fractured the Court, then, was not the protection owed to trans women, but the extent to which the tribunal was willing to extend its remedial reach beyond the specific provisions actually applied in Leonela’s case.

The question, then, is why the Court declined to attribute direct State responsibility for Leonela’s death. The Court was categorical in holding that, unlike in Vicky Hernández, it was not possible here to establish the existence of indicia of the participation of State agents, nor could the case be situated in circumstances comparable to those in which Vicky’s homicide occurred (“during a curfew, a period in which members of the security forces were the only actors present in the streets”) (fn. 140). Accordingly, although the Court accepted that Leonela’s killing took place against a proven backdrop of violence against trans women, that context alone was insufficient to ground attribution under the obligation to respect the right to life. Something more was required: evidence of “State acquiescence, tolerance, or collaboration” (para. 96). The material submitted by the Inter-American Commission and the representatives failed to demonstrate these elements.

The Court also dismissed the possibility of engaging indirect State responsibility through the duty to guarantee Leonela’s life, because it had not been established that she faced a “real and immediate” risk that was known—beyond the general context—to the authorities (para. 100). In the view of the entire bench, it was therefore not reasonable to require the State to have adopted preventive measures aimed at avoiding the materialisation of such a risk (paras. 101–102).

Moreover, at the hearing before the Court, the representatives introduced expert testimony seeking to link Leonela’s death to religiously inflected discourse by State officials at the time, calling for discrimination and violence against LGBTIQ+ persons more broadly. At the hearing, the bench repeatedly questioned how such discourse could plausibly sustain the central claim of direct or indirect attribution. Rather than strengthening the case, this move underscored the Court’s insistence on a more concrete evidentiary nexus. It is also important to recall that, in the representatives’ initial litigation strategy, Thalía was expected to testify before the Court. However, she was murdered in 2022 before the hearing could take place. Leonela’s representatives sought to substitute her testimony with that of another proposed witness, but the Court declined this request. The hearing therefore proceeded without any direct testimonial account from Leonela’s immediate circle and instead opened with expert evidence that ultimately framed the Court’s engagement with the case.

The Court draws a demanding evidentiary boundary: structural violence and context matter, but they do not, without more, substitute for proof of State involvement or specific knowledge of risk. One may reasonably wonder whether the outcome might have been different had the litigation focused exclusively on the duty to guarantee—that is, on an indirect attribution of responsibility for Leonela’s death. What is clear, in retrospect, is that the decision to pursue the more ambitious path of direct attribution—particularly through the lens of religious discourse—significantly shaped the Court’s adverse conclusion.

Conclusion

Leonela Zelaya consolidates the Court’s emerging protections for trans women, while offering a sobering reminder that evidentiary strategy may prove outcome-determinative.

At first glance, the judgment may appear to mark a retreat from the ambitious attribution of responsibility embraced by the Court in Vicky Hernández. Yet to read the judgment solely through that lens would be to miss its deeper significance. Leonela Zelaya reaffirms—with greater unity than before—that the Convention of Belém do Pará protects trans women.

What the judgment reveals is not a Court abandoning its emerging standards, but rather one signalling the limits of how far context alone can carry findings of direct State responsibility in the absence of concrete evidentiary links. For strategic litigation, the message is clear: the Inter-American system continues to advance important protections for trans women, but the path to State responsibility for transfemicide will require careful attention not only to structural patterns of violence, but also to the evidentiary architecture through which those patterns become judicially legible.

Meanwhile, the Court will soon face another crucial test. In Luiza Melinho v Brazil, it will be called upon to address the obstacles encountered by a trans woman who was prevented from accessing gender-affirming surgery, raising central questions about the scope of State obligations under the right to health. The outcome of that case will further reveal how firmly the Court is prepared to entrench—and expand—its emerging standards of protection for trans persons across the Americas.

Note: This image of Leonela and Thalía was prepared by Cattrachas, the organisation in Honduras leading the domestic and international litigation of this case, and is used with its permission. It forms part of a broader campaign aimed at raising awareness of the situation of trans women in Honduras.