Abstract Background Sacrifice of the latissimus dorsi (LD) muscle might entail donor site morbidity when used in delayed breast reconstruction. Previous studies are small, have short follow-up, and demonstrate diverging results. The aims of this study were to evaluate long-term patient-reported effects on shoulder and back function following LD flap harvest, and to investigate predictors for a worse outcome. Method This is a retrospective observational case–control cohort study. Cases were all patients who had undergone an LD flap reconstruction during the years 2007–2017. Controls were patients reconstructed with a deep inferior epigastric perforator (DIEP) flap during the same time period. Participants completed two validated questionnaires; the BREAST-Q reconstruction LD domains and the Western Ontario Shoulder Osteoarthritis Index (WOOS). Results A total of 135 cases (75 per cent) and 118 controls (60 per cent) responded to the questionnaires. The mean follow-up time was 7 years. Patients reconstructed with a LD flap were significantly less satisfied with their back and shoulder function when compared to the DIEP controls, as measured with BREAST-Q and WOOS. Predictors for a poor patient-reported back and shoulder function included axillary surgery and axillary radiotherapy, especially when combined, as well as higher age at reconstruction. Conclusion Patients who have undergone LD flap for delayed breast reconstruction had a lower satisfaction with back and shoulder function, when compared to patients who had undergone a DIEP reconstruction. Delayed LD reconstruction should be used with care, especially in patients who have undergone axillary surgery and axillary radiotherapy.