SwallPEG finds no significant swallowing difference between gastrostomy strategies in head and neck cancer

The randomized SwallPEG trial found no significant six-month swallowing quality-of-life difference between prophylactic and reactive gastrostomy, with early closure limiting certainty.

KEY POINTS

  • SwallPEG (NCT04019548) was a prospective, multicenter, randomized phase III trial comparing prophylactic versus reactive percutaneous endoscopic gastrostomy (PEG) in patients receiving definitive cisplatin-based chemoradiotherapy for locally advanced head and neck cancer. Between December 2019 and October 2025, 59 patients were randomized, with 57 eligible for analysis: 28 assigned to prophylactic PEG and 29 to reactive PEG. Oropharyngeal cancer accounted for 77.2% of the cohort.
  • Patients received 69.12 Gy in 32 fractions to high-risk volumes and 56 Gy to elective bilateral neck volumes, using IMRT/VMAT with daily image guidance. Concurrent cisplatin was administered as 100 mg/m² every three weeks or 40 mg/m² weekly. All patients received nutritional counseling, weekly dietitian assessment, and structured swallowing rehabilitation. Patients with baseline grade ≥2 dysphagia requiring dietary modification or severe malnutrition were excluded.
  • The primary endpoint, swallowing-related quality of life at six months, did not differ significantly between treatment strategies. Median MD Anderson Dysphagia Inventory (MDADI) composite scores were 69.5 with prophylactic PEG versus 74.7 with reactive PEG (p=0.595). The primary analysis included 54 patients, including prespecified proxy-imputed assessments, while complete-case sensitivity analysis included 46 patients and produced consistent findings. Swallowing scores declined during treatment and recovered toward baseline by 12–24 months in both groups.
  • The trial closed early after crossing its prespecified futility boundary. At interim analysis, mean six-month MDADI scores were 70.7 versus 71.6 (p=0.862), exceeding the predefined futility threshold of p>0.726. The original recruitment target was 110 randomized patients, but only 57 eligible patients were analyzed because recruitment was slow. The observed 5.2-point numerical difference at six months favored reactive PEG but was not statistically significant, and smaller clinically meaningful effects cannot be excluded.
  • Reactive PEG avoided tube placement in approximately half the randomized patients, although nutritional deterioration occurred in some patients who could not undergo the procedure. In the prophylactic arm, 25/28 patients underwent PEG placement, compared with 15/29 (51.7%) in the reactive arm. Of the 19 reactive-arm patients meeting insertion criteria, three declined PEG and one had a contraindication related to febrile neutropenia; all four subsequently experienced >10% weight loss during treatment. Two patients receiving prophylactic PEG never required enteral feeding.
  • Nutritional outcomes and tube dependence were comparable. Median weight change at six months was −7.0% with prophylactic versus −9.0% with reactive PEG (p=0.665). Among patients undergoing PEG placement, median time to removal was 5.10 versus 5.37 months, respectively. Enteral nutrition was initiated in 82.1% of patients randomized to prophylactic PEG versus 48.3% in the reactive group (p=0.005), reflecting differences in tube availability and use rather than demonstrated superiority in nutritional outcomes.
  • Treatment toxicity and delivery were broadly similar between arms. Grade 3 oropharyngeal mucositis occurred in 33% versus 31%, and grade 3 dysphagia in 15% versus 14%, respectively. PEG-related complications were predominantly mild, including local erythema and hypergranulation. Cisplatin was delivered without dose modification in 89% versus 83%, and all but one patient completed prescribed RT. The study lacked sufficient power for uncommon complications or oncological comparisons, and objective swallowing assessments were not included in this primary report.

CLINICAL TAKEAWAY

SwallPEG did not demonstrate a significant swallowing-related quality-of-life advantage for routine prophylactic gastrostomy over a reactive strategy during definitive head and neck chemoradiotherapy. These findings support individualized PEG placement when close nutritional surveillance and prompt intervention are available, but early closure and limited statistical power prevent a definitive conclusion of equivalence between strategies.

SOURCE

Radiotherapy and Oncology

Browse more research Suggest a correction