Classification: Documented

The Garanhuns Cannibals: three victims, two trials, and the story beyond the horror

Between Olinda and Garanhuns, Jéssica, Giselly and Alexandra were murdered from 2008 to 2012. Prosecutor records, forensic identification and court rulings trace how the three defendants were found and convicted, and why the famous pastry story requires more caution than the headlines allow.

Garanhuns, Pernambuco, Brasil · See stories from this place and nearby

The three people behind the nickname

“Garanhuns Cannibals” is the label that stuck in the press, but it directs attention to the perpetrators and the most shocking part of the case. The proceedings concern three victims first: Jéssica Camila da Silva Pereira, 17, killed in Olinda in 2008; Giselly Helena da Silva, 31; and Alexandra da Silva Falcão, 20, killed in Garanhuns in 2012. Some reports spell Giselly “Gisele” or vary her surname. This dossier follows the spelling in the Pernambuco Public Prosecutor’s Office (MPPE) bulletin.

The convicted defendants were Jorge Beltrão Negromonte da Silveira, Isabel Cristina Torreão Pires and Bruna Cristina Oliveira da Silva. This is a history of disappearances, exploited trust and killings in two cities. Claims of further victims and of pastries containing human remains carry a different evidentiary weight from the three murders tried in court. Keeping those distinctions visible prevents real women from becoming props in a crime legend.

Olinda, 2008: Jéssica and the first murder

The first killing took place in Olinda, near Recife, rather than in Garanhuns. Jéssica was a teenager and the mother of a young girl. According to MPPE’s reconstruction, she became involved with the group, was killed in 2008 and the defendants continued living with her daughter. Later accounts by the defendants offered different explanations of how she entered their home; none should be mistaken for a neutral account of the teenager’s wishes.

The surviving child adds a dimension rarely discussed without sensationalism: after her mother’s death she spent years with the people later convicted of killing Jéssica. Her name, address and present life are unnecessary to understand the case and are omitted here. When the Garanhuns investigation erupted in 2012, it led investigators back to the earlier Olinda killing. MPPE explains that the cases were split because the crimes occurred in different jurisdictions, not because there were different numbers of known victims.

Garanhuns, 2012: two approaches, two deaths

In Garanhuns, according to the indictment summarized by the Pernambuco Court of Justice (TJPE), Giselly was killed on 25 February 2012 and Alexandra on 12 March. Prosecutors said Giselly had been drawn in with a promise of spiritual guidance, while Alexandra was offered work caring for a child. The prospect of help or employment mattered: it could open the door of a home without immediately arousing suspicion. The defence was entitled to contest that account at trial; these dates are those given in the formal prosecution narrative.

The juries found the three defendants responsible for the deaths. The consulted records also describe concealment and desecration of bodies, alongside property and identity offences connected to the sequence. Reproducing the violence in graphic detail adds nothing to the documentary record. What matters to the timeline is that two women disappeared in the same city within weeks, and that proximity helped investigators connect clues that initially looked separate.

April 2012: how investigators reached the house

Investigations into Giselly’s and Alexandra’s disappearances led police to the defendants’ home. On 11 April 2012, human remains were found buried on the property and the suspects were detained. MPPE reports that information from the child living with them, together with statements from the adults, helped expose Jéssica’s earlier death in Olinda. The girl was a vulnerable witness: her statements belong within the context of investigation and protection, not entertainment.

A further allegation connects the episodes. According to TJPE, Bruna used an identity document belonging to the earlier victim when she went to a police station that day. The inquiry therefore became more than a search for two recently missing women: that clue pointed to a different identity and a different city. It did not make every subsequent statement automatically reliable, but it helps explain how an investigation in Garanhuns reached a murder four years earlier.

What DNA identified and what it did not

In May 2012, Pernambuco forensic authorities identified Giselly’s and Alexandra’s remains through DNA testing, according to reporting from the time. This is forensic evidence of the identity of the two victims found. It is not, by itself, proof that a particular pastry sold in Garanhuns contained human material. A forensic test has a defined object: identifying remains, comparing samples or examining a specific trace. Applying its result to a different object is a logical leap.

Early reporting still treated the identification of material linked to Jéssica as a separate step. Her death later formed the Olinda case and led to convictions. This timing helps explain why reports from 2012 expressed different degrees of certainty. The contemporary DNA report can be read alongside the MPPE bulletins of 2014 and TJPE rulings, rather than treating any single headline as the complete laboratory report.

A manuscript and the “sect” explanation

Investigators found a text by Jorge publicised as Revelações de um Esquizofrênico (“Revelations of a Schizophrenic”), in which he described his own beliefs and justifications. Reports also repeated a claimed mission of “purification” and the name “Cartel” for a group. The text is relevant to the defendant’s self-image and narrative. On its own it establishes neither an external organisation nor a supernatural command nor a complete psychological motive. A self-authored account can combine intention, fantasy and attempts to control the story.

Nor should the manuscript’s title be turned into a diagnosis. In the Olinda case, TJPE reported that the psychiatric assessment adopted by the court found the defendants able to understand the wrongfulness of their actions and act on that understanding. This is a legal and forensic conclusion about criminal responsibility; it does not imply that mental illness in general causes violence. The jurors evaluated actions and evidence, not the explanatory force of a provocative title.

The pies and savoury pastries: a statement is not a food test

The detail that travelled worldwide was the claim that victims’ remains had been used in savoury pastries sold to the public. There is a documentary basis for saying that police recorded such a statement from Isabel and that journalists published parts of it. But coverage of the 2014 trial, preserved in a TJPE media compilation, also reports that she denied putting human remains in the pastries and said she had made the earlier claim out of fear. Jorge likewise disputed the sales. The changing accounts need to remain labelled as accounts.

The sources reviewed for this dossier did not provide a laboratory report on pastries sold or a verified list of customers. That absence does not prove such sales impossible; it defines what these sources allow us to say. The TJPE descriptions of the convictions centre on murder, concealment and desecration of bodies, plus related offences in the Garanhuns case. “Police recorded a later contested statement” is more defensible than declaring that every customer ate human flesh as an established fact.

Three confirmed victims, not an open tally

Talk of a larger number of deaths circulated during the investigation. In April 2012, Jornal do Commercio reported that police could substantiate three killings, despite speculation about eight victims. A confession, manuscript or apparent pattern can guide inquiries. Recognising another victim requires an identifiable person, circumstances and supporting evidence. Without that bridge, a hypothesis becomes a false result through repetition.

This boundary protects the families of missing people as well. Associating someone with a notorious case without evidence can cause further public harm and divert attention from the established victims. This dossier names three women because these are the three murders in the decisions and records reviewed. Should new documents or judgments emerge, the count should be revisited on that basis rather than through lists copied across websites.

2014: the first jury, for the Olinda killing

In November 2014, an Olinda jury convicted all three defendants for Jéssica’s death. TJPE lists aggravated murder, concealment and desecration of a body among the offences. Jorge initially received 21 years and six months of imprisonment plus one year and six months of detention; Isabel and Bruna each received 19 years of imprisonment plus one year of detention. Those figures describe the original judgment, not an unchanging present sentence.

The Olinda proceedings matter because they legally established the victim who had disappeared four years before the Garanhuns discovery. MPPE and TJPE records allow readers to follow the transition from investigation to trial. The defence challenged aspects of the prosecution account and of responsibility. The jury’s verdict does not mean every detail circulated in the media, especially about food and rituals, was individually proved.

2018: the second jury, for the Garanhuns deaths

The trial for Giselly and Alexandra took place in Recife after transfer from Garanhuns amid concerns about intense local attention and jury impartiality. A November 2018 sitting was postponed; the trial began on 14 December and ended on 15 December. Thus a November listing was not the verdict, and a report on the first day of December did not yet contain the final sentence.

At this second trial Jorge was initially sentenced to 71 years, Isabel to 68 years, and Bruna to 71 years and ten months. Contemporary coverage explains that these totals included the two murders and related offences. In 2021, TJPE’s Second Criminal Chamber upheld the convictions but corrected an ambiguity in sentencing and reduced Bruna’s total to 70 years and ten months. Repeating only the 2018 headline as her current sentence misses that one-year change.

Appeals: why the numbers changed

The two cases followed separate appeal paths. In the Olinda case, TJPE’s First Criminal Chamber increased the murder sentences in 2019: Jorge’s total became 27 years of imprisonment plus one year and six months of detention; Isabel’s and Bruna’s each became 24 years and six months of imprisonment plus one year of detention. According to Jornal do Commercio’s later review, Brazil’s Superior Court of Justice revised the calculation again in 2021, reducing the terms in that case.

The 2021 review of the Garanhuns case had a different outcome: the convictions were upheld and Bruna’s sentence received a specific correction. One cannot safely add figures drawn from headlines in different years and call the result a “final sentence” or “time left to serve”. Sentence execution has its own decisions and rules. The reliable core is two jury trials concerning three murders, followed by multiple appeal stages.

What remains after the shock fades

The record establishes that Jéssica, Giselly and Alexandra were murdered; that Jorge, Isabel and Bruna were convicted in separate proceedings; that DNA testing identified remains from the two Garanhuns victims; and that later rulings changed sentence figures. Limits remain around how far pastry sales extended, alleged victims beyond the three tried, and the full truth behind the defendants’ justifications. Statements and headlines alone cannot fill those gaps.

Readers can begin with MPPE bulletins, move to TJPE notices, then compare the contemporary DNA report and the 2021 appeal coverage. That route shows something beyond the nickname: the difference between an allegation, a tested trace, a verdict and a version amplified by repetition. The three women and the surviving child deserve to be remembered with that care rather than reduced to the most marketable detail of their tragedy.

Open in Google Maps · Approximate location