Development and Internal Validation of a Nomogram for Predicting Short-Term Functional Improvement After Pharmacological Treatment in Severe symptomatic Lumbar Disc Herniation

Abstract

Abstract Background Lumbar disc herniation (LDH) is an important cause of low back pain and functional impairment. For patients with severe LDH who are not immediate candidates for surgery or prefer to delay surgery, pharmacological treatment remains a major therapeutic option; however, short-term treatment responses vary substantially among individuals. At present, individualized tools for predicting short-term functional improvement after pharmacological treatment are lacking. Therefore, this study aimed to develop and internally validate a nomogram to predict 14-day improvement in the Oswestry Disability Index (ODI) in patients with severe LDH following pharmacological treatment. Methods A total of 199 patients with MRI-confirmed severe LDH from 13 centers were included. All patients received 14 days of pharmacological treatment (Chinese patent medicine (CPM) vs non-steroidal anti-inflammatory drugs (NSAIDs)). The primary outcome was change in ODI at day 14. Candidate predictors were reduced using AIC-guided multivariable modeling, and a nomogram was developed from the final model. Secondary exploratory logistic analyses were performed using clinically relevant ODI improvement thresholds (>10, >20, and >30 points), with >10 points corresponding to the minimal clinically important difference (MCID). Internal validation was performed using bootstrap resampling. Results The optimal linear prediction model included six key variables: treatment group, sex, alkaline phosphatase (ALP), angular instability, degree of disc herniation (DDH), and hypertrophy of the ligamentum flavum (HLF). In the primary linear model, CPM treatment, female sex, higher ALP levels, and DDH-protrusion were significantly associated with lower ODI improvement (all P<0.05). In secondary threshold-based analyses, sex, ALP, and HLF were significant negative factors when ODI improvement was >30 points, while DDH-protrusion was associated with a lower likelihood of ODI improvement >20 points. Calibration plots suggested acceptable agreement between predicted and observed 14-day ODI improvement in internal bootstrap validation. Conclusions This internally validated nomogram may help estimate short-term functional improvement after pharmacological treatment in severe LDH and may assist pre-treatment risk stratification. However, given the short follow-up, limited sample size, and absence of external validation, the model should be considered preliminary and requires further validation before routine clinical use.