Malnutrition tied to poorer recovery after spinal surgery in 900-patient cohort

Patients with more severe malnutrition before elective spinal surgery experienced slower functional recovery, more complications and longer hospital stays in a retrospective cohort study of 900 adults treated at a single academic spine center in China between January 2020 and December 2024. The researchers classified 270 patients as having no malnutrition, 324 with mild malnutrition, 198 with moderate malnutrition and 108 with severe malnutrition. All groups improved after surgery, but the 12-month differences were substantial: mean Oswestry Disability Index scores were 16.8 in patients without malnutrition and 49.8 in those with severe malnutrition, with lower scores indicating less disability. These observations show an association, not proof that malnutrition caused the poorer outcomes.

Malnutrition was defined using a dual-criterion framework requiring both a phenotypic sign, such as unintentional weight loss, low body mass index or reduced skeletal muscle mass, and an etiologic factor, such as reduced food intake or systemic inflammation. The study also collected laboratory measurements, anthropometry, handgrip strength, CT-derived skeletal muscle index and composite nutritional indices including the Nutritional Risk Screening-2002, Prognostic Nutritional Index, Geriatric Nutritional Risk Index, Controlling Nutritional Status and Mini Nutritional Assessment–Short Form. These measurements showed graded differences across the four groups, supporting the internal consistency of the classification. Mean serum albumin, for example, declined from 4.08 grams per deciliter in patients without malnutrition to 2.64 in the severe group, while mean handgrip strength fell from 30.2 to 11.2 kilograms.

Functional benefit was evident in every nutritional category, but its magnitude varied. The average reduction in Oswestry Disability Index from baseline to 12 months was 35.6 points in patients without malnutrition, 33.2 with mild malnutrition, 28.0 with moderate malnutrition and 21.6 with severe malnutrition. Achievement of the study’s 12-point minimum clinically important difference was correspondingly lower: 90.4% of patients without malnutrition met the threshold, compared with 85.8%, 78.8% and 63.0% in the mild, moderate and severe groups. Twelve-month back-pain scores averaged 1.8 on a 0-to-10 scale in the no-malnutrition group but 5.4 in the severe group, while EQ-5D-5L quality-of-life utility scores averaged 0.78 and 0.38, respectively.

Differences in independence and mobility were also pronounced. Patients without malnutrition first ambulated after a mean of 1.2 days, compared with 4.2 days in the severe group. At three months, 90.4% of patients without malnutrition had returned to independent activities of daily living, versus 38.9% of those with severe malnutrition. At 12 months, 80.0% and 25.9%, respectively, had returned to their prior work or habitual activity. In an adjusted logistic model, severe malnutrition was associated with 5.64 times the odds of poor functional outcome, defined as an Oswestry Disability Index above 40 at 12 months, compared with no malnutrition; the 95% confidence interval was 3.32 to 9.58. Moderate malnutrition was associated with 3.12 times the odds and mild malnutrition with 1.68 times the odds.

Time-to-event analysis suggested that clinically meaningful recovery also occurred later in malnourished patients. In a Cox proportional-hazards model, the adjusted hazard ratio for achieving the Oswestry minimum clinically important difference was 0.46 for severe malnutrition relative to no malnutrition, with a 95% confidence interval of 0.35 to 0.60. Median time to the threshold was 3.2 months in patients without malnutrition, 4.8 months with mild malnutrition, 7.4 months with moderate malnutrition and 10.8 months with severe malnutrition. The overall adjusted model had a concordance index of 0.738. Because this was an observational analysis, the hazard ratios describe differences in recovery timing and do not establish that nutritional status alone produced the delays.

Preoperative malnutrition was accompanied by a steep gradient in perioperative complications. Any postoperative complication occurred in 10.4% of patients without malnutrition, 22.2% with mild malnutrition, 36.4% with moderate malnutrition and 59.3% with severe malnutrition. Major complications defined as Clavien-Dindo grade III or higher ranged from 3.7% in the no-malnutrition group to 24.1% in the severe group, while 30-day mortality ranged from 0% to 7.4%. Superficial surgical-site infection increased from 1.5% to 22.2%, deep infection from 0.7% to 14.8% and sepsis from 0% to 11.1% across the same comparison. Mean hospital stay rose from 4.2 to 14.6 days, intensive-care admission from 1.5% to 25.9%, and non-routine discharge from 4.4% to 63.0%.

The severity groups differed in ways that complicate causal interpretation. Patients with severe malnutrition were older, had lower body mass index, more comorbidities, poorer American Society of Anesthesiologists physical-status scores and more inflammation than those without malnutrition. The groups had similar sex distributions and broadly similar mixes of spinal diagnoses, but the differences in age and disease burden were substantial. Although the models adjusted for measured demographic, clinical and surgical factors, the study could not eliminate all unmeasured confounding or potential selection bias. It was also confined to one institution, included only patients undergoing elective surgery for degenerative disorders, and relied on routine clinical records. Nutritional intake and supplementation were not systematically reassessed throughout follow-up.

The authors also performed exploratory mediation analyses to examine possible pathways connecting malnutrition with functional outcomes. For severe malnutrition, postoperative complications, prolonged hospitalization and delayed ambulation together accounted statistically for 58.0% of the estimated association with 12-month Oswestry improvement, while the remaining 42.0% was described as a direct association not passing through those measured mediators. The largest individual indirect estimate was for postoperative complications, followed by length of hospital stay and time to first ambulation. These are model-based decompositions of an observational association, not demonstrated biological mechanisms, and the paper classified the mediation work as exploratory rather than conclusive.

For predicting poor 12-month functional outcome, a single clinical grade of moderate or severe malnutrition achieved an area under the receiver operating characteristic curve of 0.798, with 76.3% sensitivity and 78.4% specificity. Prealbumin below 20 milligrams per deciliter had an area of 0.808, and a Prognostic Nutritional Index below 40 had an area of 0.804. Combining clinical staging with laboratory and nutritional measures produced higher discrimination. A full multimarker panel had an area of 0.878, sensitivity of 84.8% and specificity of 86.2%, outperforming the grade-based classification alone. These estimates came from the same retrospective cohort, however, and require external validation before their predictive performance can be generalized to other hospitals or surgical populations.

The study also compared 315 malnourished patients who received a protocolized nutritional intervention with 315 who received standard care after propensity-score matching on age, sex, body mass index, malnutrition grade, diagnosis, surgical type and other characteristics. The protocol included high-protein oral supplements providing at least 18 grams of protein twice daily, individualized dietitian counseling, a protein target of at least 1.5 grams per kilogram per day, and support extending beyond discharge. Matched patients receiving the protocol had a mean 12-month Oswestry score of 27.4 versus 36.8 in the standard-care group, and 85.1% versus 66.0% achieved the minimum clinically important difference. Any postoperative complication occurred in 21.6% and 38.7%, respectively, and mean hospital stay was 3.4 days shorter in the intervention group. These results are promising but remain observational: treatment was not randomized, matching can balance measured characteristics only, and differences in adherence or clinical management may have contributed. The findings therefore support further controlled evaluation of perioperative nutritional support rather than establishing it as a guaranteed cause of better recovery.

Subject of Research: Preoperative malnutrition severity and functional recovery after elective spinal surgery

Article Title: Association of nutritional risk with postoperative functional outcomes in patients undergoing spinal surgery: a retrospective cohort study

Article References: Chen, D., Yang, Y., Guo, Y., & Zhu, Z. (2026). Association of nutritional risk with postoperative functional outcomes in patients undergoing spinal surgery: a retrospective cohort study. Frontiers in Nutrition, 13, Article 1926068. https://doi.org/10.3389/fnut.2026.1926068

Image Credits: AI Generated

DOI: 10.3389/fnut.2026.1926068

Keywords: malnutrition, spinal surgery, functional recovery, postoperative complications, nutritional assessment, Oswestry Disability Index, quality of life, clinical nutrition, retrospective cohort, nutritional support, Association, nutritional