KOHEN AVOCATS FIRM · PARIS 16
Wanted by Ukraine? Why Paris, Rome and Rotterdam Say No While London Says Yes

Hassan KOHEN acts in urgent matters, from the police station to the assize court. Initial strategic analysis: €80 incl. VAT, personal reply within 24 hours.

100% confidential · Legal professional privilege · No obligation

Article generated by artificial intelligence, through a process designed and supervised by the firm

Source: French Court of Cassation – Judilibre and Légifrance open data

Paris Bar Police custody, investigation, assize court CNB listing avocat.fr
Maître Hassan KOHEN, avocat au Barreau de Paris
Hassan KOHEN
Attorney at the Paris Bar

Wanted by Ukraine? Why Paris, Rome and Rotterdam Say No While London Says Yes

Speak to a lawyer 06 89 11 34 45

On 29 October 2025, the Ninth Investigating Chamber of the Paris Court of Appeal refused to surrender a man wanted by Ukraine. Every formal box had been ticked: the request was regular, the facts were punishable in both countries, and Kyiv had supplied written assurances that his rights would be respected. The court set those assurances aside anyway. Martial law, it held, covers the whole of Ukrainian territory rather than a defined combat zone, the derogations Kyiv lodged to the European Convention on Human Rights are too broad and open-ended to be tested, and no paper promise can cure a risk the requesting state is objectively unable to prevent. Four months later, on 4 March 2026, the Westminster Magistrates’ Court in London looked at the same war, the same martial law and the same type of Ukrainian assurances — and reached the opposite conclusion, accepting the guarantees and ordering the extradition process forward. Between those two poles, Italy’s Court of Cassation and the Rotterdam District Court have built, since late 2025, a demanding European mainstream: surrender to an active war zone requires more than formal guarantees; it requires proof that protection works in practice. For anyone sought by Ukraine and arrested in France, that split is not academic. It dictates which arguments win before the Paris chambre de l’instruction, which evidence to file, and when an appeal to London’s reasoning by the prosecution must be answered with Paris’s. This analysis sets out what each court decided, what the French Court of Cassation already requires in this exact type of case, and how a defence team in Paris turns the divergence into a refusal.

I. Why Do French, Italian and Dutch Courts Refuse Surrender to Ukraine?

A. How did the Paris Court of Appeal dismantle Ukraine’s assurances on 29 October 2025?

The Paris judgment, handed down by the Ninth Investigating Chamber under case number 2024/10328, matters first for its method. The court did not dispute that the basic conditions of extradition were met, and it did not question Ukraine’s good faith in the abstract. It examined the content of the guarantees against the reality they were supposed to cover — and found the two did not match. The decisive passage of the reasoning, as reported with the full text of the decision, turns on the scope of martial law: the court rejected the idea that emergency measures were confined to active combat zones and held instead that martial law extends across the entire Ukrainian territory, so that assurances calibrated for ordinary criminal justice cannot answer risks created by a nationwide wartime regime. The fluid and unpredictable nature of hostilities, including strikes reaching far behind the front line and hitting civilian areas, means the danger follows the person wherever the requesting state might detain or try him.

The second pillar of the Paris reasoning concerns the derogations Ukraine formally lodged to the European Convention on Human Rights and to the International Covenant on Civil and Political Rights in 2022, extending them ever since. Derogating in wartime is lawful in itself. What the court found fatal is that these derogations are, for surrender purposes, neither sufficiently limited nor clearly circumscribed: a judge asked to authorise a surrender cannot verify which rights remain enforceable, where, and for whom, when the derogating state itself reserves sweeping emergency powers. A guarantee that fair-trial rights will be respected loses its evidentiary value when the guarantor has simultaneously suspended, across the whole territory, the very framework that would make the promise verifiable. In practical terms, the Paris court raised the bar from formal sufficiency — a sealed assurance on official letterhead — to functional sufficiency: can this state, here and now, actually deliver what it promises?

That shift mirrors exactly what the French Court of Cassation did for Russia after 16 September 2022, when the Russian Federation ceased to be a party to the Convention. In a decision of 13 December 2022 (no. 22-80.610), the Criminal Chamber quashed a surrender opinion that had relied on Moscow’s solemn undertaking to respect Convention rights, holding that once Russia had left the Convention system, an opinion relying on Moscow’s undertaking was deprived of legal foundation: “en ce qu’il est fondé sur l’engagement solennel des autorités russes de respecter les droits et libertés fondamentaux garantis par la Convention européenne des droits de l’homme, l’arrêt, qui se trouve privé de fondement juridique, ne satisfait pas aux conditions essentielles de son existence légale, au regard des articles 1er de la Convention européenne des droits de l’homme et 696-15 du code de procédure pénale”. The full decision is published by the Court of Cassation at Cass. crim., 13 December 2022, no. 22-80.610. The parallel with Ukraine is direct and should be pleaded expressly: where the requesting state’s relationship with the Convention framework has been emptied of verifiable substance — by exit in the Russian case, by sweeping wartime derogation in the Ukrainian one — assurances denominated in Convention currency no longer purchase a lawful surrender. Defence teams should file the derogation texts themselves, with their renewal history, and force the prosecution to explain which Convention rights remain concretely enforceable in the region of detention, rather than accepting general references to the Ukrainian Constitution or to ministerial letters.

The Paris approach also draws strength from the Criminal Chamber’s death-penalty and inhuman-treatment case law, which already tells investigating chambers what to do when the risk is real. On 30 January 2024 (no. 23-83.549), the Court held: “Le premier de ces textes prohibe l’extradition vers un autre État d’une personne dont il y a des motifs sérieux et avérés de croire qu’elle serait exposée dans le pays de destination à un risque réel d’être soumise à la peine de mort.” The decision is available at Cass. crim., 30 January 2024, no. 23-83.549. And on 13 January 2021 (no. 20-81.359), it ruled: “Il résulte de ces textes que la chambre de l’instruction qui constate que la personne réclamée encourt, en cas d’extradition vers son pays d’origine, le risque d’être soumise à un traitement inhumain et dégradant, doit donner un avis défavorable.” The Court added the operational trigger: “Un tel risque est avéré lorsque la personne bénéficie de la protection subsidiaire aussi longtemps qu’il n’y a pas été mis fin.” See Cass. crim., 13 January 2021, no. 20-81.359. Read together, these rulings convert the Paris Ukraine analysis into a general method: identify the concrete risk (death penalty, inhuman detention, unfair trial under martial law), test the assurance against the state’s objective capacity to honour it, and where the risk survives the paper, the opinion must be unfavourable. A Ukrainian file combining conscription-related charges, detention in wartime facilities and trial before courts operating under martial law fits that method better than almost any other current caseload.

B. What do Rome and Rotterdam add: the Italian Cassation rulings and the air-raid shelters test?

Italy has turned the same intuition into supreme-court doctrine. In two recent rulings — Section VI, no. 4828 of 12 November 2025 (published 5 February 2026) and no. 1167 of 30 December 2025 (published 13 January 2026) — the Italian Court of Cassation held that general assurances offered by Ukraine are increasingly insufficient to dispel the risk, and that extradition must consequently be refused. The decisive factor, in the Italian formulation, is no longer the formal existence of safeguards but their practical reliability in an evolving war context: where the requesting state is objectively unable to ensure that the individual will not face a real and current risk to life, physical integrity or fundamental rights, surrender must go. That is the functional-sufficiency test in its purest form, and it is now the law in two of the European Union’s three largest criminal-justice systems. A Paris prosecutor who argues that the October 2025 refusal was an isolated first-instance assessment must answer Rome: the highest Italian court independently reached the same result, twice, within weeks.

Rotterdam adds the most granular factual template. On 13 October 2025 and again on 3 March 2026, the Rotterdam District Court denied extradition to Ukraine, grounding the refusals in Articles 2, 3 and 13 of the Convention and relying on the war’s concrete effects on detention conditions, the risk of air raids, the lack of adequate shelters in places of detention, and the absence of an effective remedy. This is the level of factual detail that wins cases: not “there is a war”, but photographic and expert evidence of the specific detention centre’s shelters, the frequency of alerts in its district, and the documented impossibility of an effective complaint while the emergency regime lasts. Note the nuance the Dutch judges preserved — the October 2025 judgment expressly dismissed the requested person’s Article 6 fair-trial challenge while granting refusal on Articles 2, 3 and 13, and the March 2026 judgment did not rule on Article 6 at all. Defence teams should learn from that discipline: load the file with winnable, evidence-backed risks (detention conditions, physical safety, remedy) rather than betting everything on a systemic fair-trial objection a court may find premature. Our companion analysis of three courts answering the same German warrants differently in the Nord Stream affair shows the same lesson from the opposite direction: surrender judges decide on concrete, individualised risk files, not on geopolitics.

Taken together, Paris, Rome and Rotterdam describe a converging European mainstream. The obstacle to surrender is not the mere existence of an armed conflict — courts have surrendered to post-conflict states before — but the widespread and unpredictable character of this one, combined with derogations that remove the yardstick by which assurances are measured. General promises are discounted; practical reliability is priced in. That convergence is the single strongest argument available in a French investigating chamber today, because it lets counsel present the Paris 2025 refusal not as national exceptionalism but as the French expression of a shared European standard.

II. How Do You Fight a Ukrainian Extradition Request From France — and Why Does London Disagree?

A. What must a Paris chambre de l’instruction verify before giving its opinion?

French extradition procedure outside the European Arrest Warrant is codified, adversarial and multi-stage, and each stage supplies a defence lever. The starting point is statutory: “En l’absence de convention internationale en stipulant autrement, les conditions, la procédure et les effets de l’extradition sont déterminés par les dispositions du présent chapitre.” See article 696 of the Code of Criminal Procedure. France and Ukraine cooperate under the 1957 European Convention on Extradition, so the treaty applies first — but the Code fills every gap, and French courts apply its protective provisions as the floor below which no surrender can fall.

The absolute bars come next, and counsel should test each one before touching the merits of the risk. Article 696-4 of the Code of Criminal Procedure provides: “L’extradition n’est pas accordée : 1° Lorsque la personne réclamée a la nationalité française, cette dernière étant appréciée à l’époque de l’infraction pour laquelle l’extradition est requise ; 2° Lorsque le crime ou le délit à un caractère politique ou lorsqu’il résulte des circonstances que l’extradition est demandée dans un but politique”. See article 696-4 of the Code of Criminal Procedure. In Ukrainian files, the political-offence and political-purpose exceptions deserve serious screening: prosecutions connected with draft evasion, desertion, collaboration accusations in occupied territories or dissent against the wartime authorities can, depending on the file, take on a political character or reveal a political motive that the requesting documents seek to dress in ordinary criminal law. French nationality at the time of the offence ends the case outright. Territoriality (offences committed in France, or already finally judged here) does the same. These bars are examined before any balancing: if one fits, the chamber never reaches the assurances debate.

If no absolute bar applies, the procedure moves to the investigating chamber, and the deadlines are short. Article 696-15 provides: “Lorsque la personne réclamée a déclaré au procureur général ne pas consentir à son extradition, la chambre de l’instruction est saisie, sans délai, de la procédure.” The person appears “dans un délai de dix jours ouvrables à compter de la date de sa présentation au procureur général”, and if he confirms his refusal, “la chambre de l’instruction donne son avis motivé sur la demande d’extradition”. See article 696-15 of the Code of Criminal Procedure. Two practical consequences follow. First, the defence file — derogation texts, detention-condition evidence, medical certificates, proof of subsidiary protection or asylum-seeker status — must be substantially ready within days of arrest, not weeks; the chamber can order additional investigation, as it did in the Tunisian death-penalty case by demanding a note verbale on non-execution, but counsel cannot count on it. Second, the opinion must be reasoned, and the Court of Cassation enforces that strictly: insufficient or contradictory reasoning equals absence of reasoning and sends the case back.

Two further provisions shape the hearing itself. The requesting state may be authorised to intervene at the hearing through a designated representative — “La chambre de l’instruction peut, par une décision qui n’est susceptible d’aucun recours, autoriser l’Etat requérant à intervenir à l’audience au cours de laquelle la demande d’extradition est examinée” — but “il ne devient pas partie à la procédure”. See article 696-16 of the Code of Criminal Procedure. Expect Ukraine to use that seat to renew its assurances orally; prepare to cross-examine them against the derogation record and the Rotterdam-style detention facts. And liberty must be fought for in parallel: “La mise en liberté peut être demandée à tout moment à la chambre de l’instruction”, which rules within twenty days at the latest. See article 696-19 of the Code of Criminal Procedure. In a file where surrender itself is contested on Article 3 and Article 6 grounds, prolonged pre-surrender detention is both a human cost and a litigation asset, and release applications keep the pressure on a procedure the defence wants slowed down.

If the opinion is favourable, the case enters its governmental phase: “l’extradition est autorisée par décret du Premier ministre pris sur le rapport du ministre de la justice”, and if the person has not been handed over within one month of notification of the decree to the requesting state, he is released as of right and can no longer be claimed on the same grounds. See article 696-18 of the Code of Criminal Procedure. The decree is subject to judicial review for excess of power before the Council of State, where the Convention arguments — Articles 2, 3, 6 and 13 — are pleaded again, this time against the government itself. The Paris, Rome and Rotterdam refusals, the Cassation trilogy on assurances, death penalty and subsidiary protection, and the derogation record together make that review a genuine second trial of the risk, not a formality. Counsel should therefore build the chamber file from day one as if the Council of State will read it, because if the opinion goes the wrong way, it will.

B. Why did Westminster say yes where Paris said no — and how do you exploit the split?

The London counterpoint is the judgment delivered by the Westminster Magistrates’ Court on 4 March 2026 in the Dmytruk case. The court refused extradition to Ukraine, but on Article 8 grounds — the right to respect for private and family life — while discarding the existence of a risk of an unfair trial under Articles 5 and 6. Its reasoning rested on two propositions Paris had rejected: that the relevant martial-law provisions do not apply where ordinary law continues to operate, as is the case in western Ukraine, and that even if some residual risk of prejudice or unfairness remained, the assurances supplied by the Ukrainian authorities provided concrete guarantees. In short, London territorialised the risk and credited the paper; Paris nationalised the risk and tested the paper against the state’s capacity. Both courts looked at the same war. They applied different theories of what an assurance is worth when the guarantor governs under emergency powers.

For a defence team in Paris, the Westminster judgment is a gift disguised as a threat. The threat is obvious: the prosecution will cite it to argue that a fellow European court, applying the same Convention, found Ukrainian assurances sufficient — so the Paris refusal should be treated as an outlier and not followed. The gift is structural. First, the London court still refused surrender, albeit on Article 8: even the most assurance-friendly court in the sample did not send the man back. Shield your client with every applicable ground, and the assurance debate alone does not decide the case. Second, the territorialisation thesis — ordinary law in the west, emergency only in the east — is factually contestable and legally fragile: missiles have struck Lviv, Uzhhorod’s region hosts strained detention infrastructure, and the derogations themselves are drawn nationwide, not province by province. File air-alert records for the specific oblast of prospective detention, expert evidence on the nationwide scope of the emergency legislation, and the derogation instruments showing no geographic limitation, and the London distinction collapses into the Paris finding. Third, the “concrete guarantees” credited in London were never stress-tested against non-compliance: ask what remedy exists if Ukraine breaches them mid-detention, and the answer runs into the Rotterdam finding — no effective remedy under Article 13 while the emergency lasts.

The deeper point is one of litigation framing. When two European courts split on the same requesting state, the French investigating chamber is not bound by either foreign decision, but it must show its work under Article 696-15’s reasoned-opinion requirement and under the Convention as interpreted by the Court of Cassation. The Cassation’s own trajectory — from the 2021 subsidiary-protection rule, through the 2022 Russia-assurances quashing, to the 2024 death-penalty prohibition — points one way: toward verification, not deference. Present the split as a choice between a verification method (Paris, Rome, Rotterdam, and the Cassation’s own logic) and a deference method (London), and invite the chamber to locate itself in the majority. Prosecutors who invoke London must then explain why the chamber should prefer the minority method that no supreme court on the continent has endorsed — Italy’s Cassation having twice gone the other way. Add that the London court’s own Article 8 refusal shows even deference has limits, and the prosecution’s best foreign authority becomes a witness for the defence.

There is also a forward-looking use of the split. Files move: a person refused in France may be sought elsewhere, and a person surrendered from elsewhere may surface in France. Document the divergence in the French file so it travels — a reasoned Paris refusal citing Rome and Rotterdam becomes persuasive material before any other European surrender judge, just as the Westminster reasoning must be rebutted wherever it appears. Where the client has family, residence or proceedings in several states, coordinate the filings so the verification method is pleaded consistently in each forum; contradictory defence theories across borders destroy credibility faster than any prosecution argument. And where the Ukrainian request concerns facts that could also interest the French universal-jurisdiction prosecutor — war crimes and crimes against humanity committed abroad — remember that a French prosecution, if opened, can complicate or block surrender dynamics entirely, a lever the London forum does not offer in the same form.

Conclusion

The map of Ukrainian extradition in Europe now has a clear shape. Paris refuses where assurances cannot survive contact with nationwide martial law and open-ended derogations. Rome refuses twice at supreme-court level where guarantees lack practical reliability in an evolving war. Rotterdam refuses where detention centres lack shelters and remedies lack effectiveness. London alone credits the paper — and even London would not send its man back, stopping the surrender on Article 8. For the person sought by Kyiv and arrested in France, the practical conclusion is direct: contest early, contest concretely, and contest comparatively. Test the absolute bars of article 696-4 first. Build the risk file within the ten working days of article 696-15 around derogation texts, detention facts and medical evidence rather than generalities about war. Plead the Cassation trilogy — subsidiary protection as proven risk, Russian assurances as voided paper, death-penalty risk as an absolute prohibition — as the domestic spine of the argument. Set the Paris refusal of 29 October 2025 inside the Rome-Rotterdam mainstream so it reads as European orthodoxy, and dismantle the Westminster exception province by province, alert log by alert log. Assurances are evidence, not verdicts; in wartime, the requesting state must prove they work, not merely that they exist. That is what Paris held, what Rome confirmed, and what any chamber asked to surrender a man to a country under martial law should hold next.

📄 Official circular