STUDY QUESTION What is the influence of age and chemotherapy regimen on the longitudinal blood anti-Müllerian hormone (AMH) variations in a large series of adolescents and young adult (AYA) (15–24 years old) and non-AYA (25–35 years old) lymphoma patients? SUMMARY ANSWER In case of alkylating regimen treatment, there was a deep and sustained follicular depletion in AYA as well as non-AYA patients; however in both groups, the ovarian toxicity was extremely low in cases of non-alkylating treatments. WHAT IS KNOWN ALREADY AMH is now well-recognised to be a real-time indicator of ovarian follicular depletion and recovery in women treated by chemotherapy. Its longitudinal variations may discriminate between highly and minimally toxic protocols regarding ovarian function. It has been shown, in different cancer types, that age, type of chemotherapy regimen and pre-treatment AMH levels are the main predictors of ovarian recovery. Large studies on longitudinal AMH variations under chemotherapy in lymphoma patients are few but can provide the opportunity to assess the degree of follicle loss at a young age. STUDY DESIGN, SIZE, DURATION This prospective cohort study was conducted in the Fertility Observatory of the Lille University Hospital. Data were collected between 2007 and 2016. Non-Hodgkin or Hodgkin lymphoma patients (n = 122) between 15 and 35 years old were prospectively recruited before commencing chemotherapy. Patients were treated either by a non-alkylating protocol (ABVD group; n = 67) or by an alkylating regimen (alkylating group; n = 55). PARTICIPANTS/MATERIALS, SETTING, METHODS Serial AMH measurements were performed at baseline (AMH0), 15 days after the start of chemotherapy (AMH1), 15 days before the last chemotherapy cycle (AMH2), and at time 3, 6, 9, 12, 18 and 24 months from the end of chemotherapy. The whole study population was divided into two groups according to age: AYA (15–24; n = 65) and non-AYA (25–35; n = 57). All patients received a once monthly GnRH agonist injection during the whole treatment period. A linear mixed model was used to account for the repeated measures of single patients. MAIN RESULTS AND THE ROLE OF CHANCE At baseline, non-AYA patients had higher BMI and lower AMH levels than AYA patients. All AYA and non-AYA patients having received ABVD protocols had regular cycles at 12 months of follow-up. In case of alkylating regimens, amenorrhoea was more frequent in non-AYA patients than in AYA patients at 12 months (37% vs 4%, P = 0.011) and at 24 months (24% vs 4%, P = 0.045). We distinguished a similar depletion phase from AMH0 to AMH2 between ABVD and alkylating groups but significantly different recovery phases from AMH2 to AMH + 24 months. AMH recovery was fast and complete in case of ABVD protocols whatever the age: AMH reached pre-treatment values as soon as the 6th month of follow-up in the AYA group (mean (95% CI) in log AMH M0 vs M6: 3.07 (2.86 to 3.27) vs 3.05 (2.78 to 3.31), P = 1.00) and in the non-AYA group (mean (95% CI) in log AMH M0 vs M6: 2.73 (2.40 to 3.05) vs 2.47 (2.21 to 2.74), P = 1.00). In contrast, no patients from the alkylating group returned to pre-treatment AMH values whatever the age of patients (AYA or non-AYA). Moreover, none of the AMH values post-chemotherapy in the non-AYA group were significantly different from AMH2. Conversely in the AYA group, AMH levels from 6 months (mean (95% CI) in log AMH: 1.79 (1.47 to 2.11), P < 0.001) to 24 months (mean (95% CI) in log AMH: 2.16 (1.80 to 2.52), P ≤ 0.001) were significantly higher than AMH2 (mean (95% CI) in log AMH: 1.13 (0.89 to 1.38)). Considering the whole study population (AYA and non-AYA), pre-treatment AMH levels influenced the pattern of the AMH variation both in alkylating and ABVD protocols (interaction P-value = 0.005 and 0.043, respectively). Likewise, age was significantly associated with the pattern of the recovery phase but only in the alkylating group (interaction P-value =0.001). BMI had no influence on the AMH recovery phase whatever the protocol (interaction P-value = 0.98 in alkylating group, 0.72 in ABVD group). LIMITATIONS, REASONS FOR CAUTION There was a large disparity in subtypes of protocols in the alkylating group. The average duration of chemotherapy for patients treated with alkylating protocols was longer than that for patients treated with ABVD. WIDER IMPLICATIONS OF THE FINDINGS These results make it possible to develop strategies for fertility preservation according to age and type of protocol in a large series of young lymphoma patients. In addition, it was confirmed that young age does not protect against ovarian damage caused by alkylating agents. STUDY FUNDING/COMPETING INTEREST(S) This work was supported by Agence Régionale de Santé Hauts de France and Agence Onco Hauts-de-France who provided finances for AMH dosages (n° DOS/SDES/AR/FIR/2019/282). There are no competing interests