Patients at the lowest risk of receiving a percutaneous endoscopic gastrostomy tube had a body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) of 25 or higher and were treated with conventional radiation therapy, while patients treated with accelerated irradiation fractionation were at the highest risk of reactively requiring the placement of a percutaneous endoscopic gastrostomy tube.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Strom T, Trotti AM, Kish J, et al. Risk Factors for Percutaneous Endoscopic Gastrostomy Tube Placement During Chemoradiotherapy for Oropharyngeal Cancer. JAMA Otolaryngol Head Neck Surg. 2013;139(11):1242–1246. doi:10.1001/jamaoto.2013.5193
Patients with oropharyngeal squamous cell carcinoma undergoing chemoradiotherapy may require percutaneous endoscopic gastrostomy (PEG) tube placement because of dehydration or significant weight loss.
To determine the need for the reactive placement of a PEG tube during chemoradiotherapy for oropharyngeal cancer and to identify patient or tumor factors associated with reactively requiring the placement of a PEG tube.
Design, Setting, and Participants
Single-institution retrospective review of 297 patients treated with intensity-modulated radiation therapy and concurrent chemotherapy for oropharyngeal squamous cell carcinoma between May 1, 2004, and June 30, 2012, with a minimum follow-up period of 3 months.
Placement of a PEG tube.
Main Outcomes and Measures
Logistic regression analysis was used to identify independent risk factors associated with symptomatic requirement for the reactive placement of a PEG tube.
In total, 128 patients did not receive a prophylactic PEG tube within 10 days of initiation of chemoradiotherapy. Fifteen of 128 patients (11.7%) required the reactive placement of a PEG tube during or within 3 months of chemoradiotherapy. The median time to PEG tube removal was 3.3 months, and 14 of 15 patients had their PEG tube removed at the last follow-up analysis. Independent risk factors for PEG tube placement included the following: accelerated irradiation fractionation (odds ratio, 4.3; 95% CI, 1.1-16.5; P = .04), a tumor T classification of 3 or higher (odds ratio, 3.5; 95% CI, 1.0-11.9; P = .04), a cumulative cisplatin dose of 200 mg/m2 or higher (odds ratio, 6.7; 95% CI, 1.2-36.7; P = .03), and a body mass index (calculated as weight in kilograms divided by height in meters squared) of less than 25 (odds ratio, 5.8; 95% CI, 1.4-23.9; P = .02).
Conclusions and Relevance
Although the overall risk is low, a body mass index of less than 25, accelerated irradiation fractionation, a tumor T classification of 3 or higher, and a cumulative cisplatin dose of 200 mg/m2 or higher are associated with symptomatic need for the reactive placement of a PEG tube in patients with oropharyngeal cancer.
Percutaneous endoscopic gastrostomy (PEG) tubes may be used as a prophylactic measure to prevent weight loss, poor nutrition, and dehydration in patients undergoing chemoradiotherapy (CRT) for head and neck cancer. The placement of prophylactic PEG tubes has been reported in several recent retrospective series as a safe and effective method for preventing nutritional decline during treatment for head and neck cancer.1-5 However, the placement of a prophylactic PEG tube carries measurable risks, most notably an infection rate of 5% to 8%, along with diarrhea, constipation, electrolyte abnormalities, gastrointestinal bleeding, a clogged lumen and tube replacement, and metastatic seeding at the gastrostomy site.4-7 In addition, some groups have suggested that avoidance of PEG tube placement during therapy may be associated with a lower risk of long-term dysphagia outcomes such as PEG tube dependence or pharyngoesophageal stricture.8-10 Therefore, our objective was to assess potential risk factors leading to the reactive placement of a PEG tube in patients undergoing concurrent CRT for oropharyngeal squamous cell carcinoma.
After institutional review board approval, 430 patients were identified who received CRT for oropharyngeal squamous cell carcinoma between May 1, 2004, and June 30, 2012. Exclusion criteria for the study included prior head and neck surgery or previous head and neck cancer, as well as induction chemotherapy, a synchronous primary tumor, or locoregional recurrence or persistence of disease within 3 months of completing CRT. From this group, 297 patients were included who were treated with intensity-modulated radiation therapy and concurrent chemotherapy and had a minimum follow-up period of 3 months. Patients who did not receive a prophylactic PEG tube before or within 10 days of initiation of CRT (unless an attempt to avoid upfront placement of a PEG tube was explicitly indicated in the medical record) were identified. A 10-day cutoff following CRT commencement to define prophylactic PEG tube placement was empirically chosen a priori because most patients do not develop treatment-related dysphagia within this period. Patient tumor and treatment details were abstracted from the medical record. The primary end points were risk factors for the reactive placement of a feeding tube. PEG tubes were reactively placed because of unacceptable weight loss or dehydration as determined by the treating physician (A.M.T., J.K., N.G.R., J.M., T.A.P., J.J.C.). The secondary end points were percentage weight loss at 3 to 4 weeks after radiation therapy commencement, at radiation therapy completion, and at 1 and 3 months following radiation therapy.
All patients were treated with intensity-modulated radiation therapy to a total dose of 62 to 70 Gy in 200-centigray daily fractions. Most patients received once-daily radiation therapy, while fewer patients underwent an accelerated irradiation regimen, consisting of an additional 200-centigray fraction each week beginning the second week, given as a twice-daily treatment with fractions separated by 6 hours or as an additional fraction on the weekend, as described by Overgaard et al.11
All patients underwent concurrent systemic therapy. Most received cisplatin every 3 weeks for 2 to 3 cycles at 75 to 100 mg/m2. The remainder were treated with one of the following regimens: weekly cetuximab for 5 to 9 cycles, concurrent weekly cisplatin for 4 to 8 cycles at 25 to 35 mg/m2, weekly carboplatin for 6 to 8 cycles with an area under the curve of 1.5 to 2.0, or carboplatin cycled every 3 weeks for 3 cycles with an area under the curve of 5.0 to 6.0.
Statistical analysis was performed using commercially available software (Statistical Product and Service Solutions, version 21.0; SPSS Inc). Wilcoxon rank sum test was used to compare continuous variables, and Pearson χ2 test with the exact method and Monte Carlo estimation was used to compare categorical variables. Multivariate analysis was performed using a logistic regression model on potential predictors from the univariate analysis. Bivariate clinical, tumor, and treatment variables were added to the logistic regression model if their univariate significance was P < .20. The same set of candidate factors was explored using conditional inference trees for identifying different subgroups of patients at risk for the reactive placement of a PEG tube.12 The factor selection was based on the univariate test, with removal at P = .05. The tree model was performed using the party R package (http://cran.r-project.org/web/packages/party/index.html). To compare relative risk, odds ratios (ORs) for the identified subgroups were generated using the logistic model. Continuous variables were split using clinically meaningful cut points. An α (type I) error of less than 0.05 was considered statistically significant.
Of 297 patients who met the inclusion criteria, 128 (43.1%) did not receive a prophylactic PEG tube within 10 days of initiation of CRT. Patient, tumor, and treatment characteristics are summarized in Table 1. Fifteen of 128 patients (11.7%) required the reactive placement of a PEG tube during or within 3 months of CRT. Among the patients who received a PEG tube during their treatment, the median time from the commencement of radiation therapy to PEG tube placement was 35 days, and the median time to PEG tube removal was 3.3 months. At the last follow-up analysis, 14 of 15 patients had their PEG tube removed, and the final patient was lost to follow-up data at 3.4 months. Compared with the patients who did not undergo PEG tube placement, the 15 patients who required PEG tubes during treatment had a significantly higher percentage median weight loss at 3 to 4 weeks after CRT commencement (6.4% vs 3.9%, P = .02) (Table 2) and immediately following CRT treatment (10.7% vs 9.3%, P = .04). However, no significant difference in weight loss was observed between the 2 groups by 1 month (12.3% vs 11.0%, P = .79) and 3 months (12.4% vs 12.5%, P = .86) following CRT.
Patient selection bias was ascertained. Compared with patients who received a prophylactic PEG tube, fewer patients who did not receive a prophylactic PEG tube had a greater than 5% pretreatment weight loss (5.5% vs 18.3%, P = .02), an American Joint Committee on Cancer T3 or T4 tumor (25.8% vs 53.3%, P < .001), and a Karnofsky Performance Scale score of less than 90 (8.6% vs 26.6%, P < .001) (range, 0-100, with 100 indicating completely normal functioning).
Risk factors for the placement of a PEG tube are given in Table 1. On univariate analysis, a cumulative cisplatin dose of 200 mg/m2 or higher (P = .05) and accelerated irradiation fractionation (P = .04) were significantly associated with the reactive placement of a PEG tube during treatment. On multivariate analysis, the following were significantly associated with the reactive placement of a PEG tube: a tumor T classification of 3 or higher (OR, 3.5; 95% CI, 1.0-11.9; P = .04), an accelerated irradiation fractionation (OR, 4.3; 95% CI, 1.1-16.5; P = .04), a cumulative cisplatin dose of 200 mg/m2 or higher (OR, 6.7; 95% CI, 1.2-36.7; P = .03), and a body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) of less than 25 (OR, 5.8; 95% CI, 1.4-23.9; P = .02). A conditional inference tree was created using the same set of potential risk factors for PEG tube placement (Figure). Patients at the lowest risk of receiving a PEG tube were those with a BMI of 25 or higher who were treated with nonaccelerated irradiation fractionation (5.6% crude rate). Conversely, patients at the highest risk of receiving a PEG tube were those who were treated with accelerated irradiation fractionation (27.8% crude rate) and those with a BMI of less than 25 who received nonaccelerated irradiation fractionation (23.8% crude rate). Relative to the lowest-risk subgroup, patients who received accelerated irradiation fractionation were 5.3 (95% CI, 1.4-20.2) times as likely to undergo the reactive placement of a PEG tube, and patients with a BMI of less than 25 who did not receive accelerated irradiation fractionation were 6.4 (95% CI, 1.6-25.4) times as likely to reactively require a PEG tube.
Among the subset of patients who did not receive a prophylactic PEG tube at our institution during CRT for oropharyngeal cancer, the results suggest that specific risk factors can be used to assess whether a patient may require a PEG tube at some point during his or her treatment. We also demonstrate little harm in postponing PEG tube placement in this patient population. The median time to PEG tube removal was 3.3 months, which is consistent with a recent large study5 in a similar patient population who underwent concurrent chemotherapy and intensity-modulated radiation therapy with prophylactic PEG tubes placed. In addition, by 1 and 3 months following CRT, no difference was observed in percentage median weight loss between those who required a PEG tube in our study and those who did not. This suggests that the use of a reactively placed PEG tube rapidly corrected their increased weight loss back to that of their peers.
Previous studies13,14 have addressed risk factors for PEG tube placement after the commencement of radiation therapy. Al-Othman et al13 performed a retrospective analysis of 934 patients with head and neck cancer and found that PEG tubes placed for acute toxic effects were associated with older age, adjuvant chemotherapy, the presence of neck disease, and a higher radiation therapy dose. Mangar et al14 found that reactively required PEG tubes were placed with increasing frequency in patients with stage III to stage IV disease, in patients actively smoking more than 1 pack per day, and in patients with a performance status of 2 to 3 (range, 0-5, with lower numbers indicating better performance). These studies add to the growing body of literature identifying potential risk factors for PEG tube placement; however, only 2 patients in these studies underwent concurrent CRT.
Most recently, Bhayani et al15 reviewed 474 patients who underwent radiation therapy with or without induction chemotherapy or CRT with or without induction chemotherapy for oropharyngeal carcinoma. The study addressed risk factors for PEG tube placement in the combined patient populations before, during, and after radiation therapy. Similar to the present study, they found that tumor T classification of 3 or 4 and accelerated irradiation fractionation by concomitant boost were predictive of PEG tube placement on univariate analysis; however, neither factor was significant on multivariate analysis. Ultimately, CRT and nonadherence to swallowing exercises were found to be independently predictive of PEG tube placement. Our study assessed only patients who underwent CRT, which precluded us from addressing this as a risk factor for PEG tube placement. In addition, we did not address patients with prophylactic receipt of PEG tubes or adherence to swallowing exercises, which likely contributed to our different findings.
As noted by Locher et al,6 confounding and selection bias are problematic in many studies addressing variables associated with PEG tube placement and PEG tube dependence. The decision to place a tube before and during treatment is a subjective decision made by the physician. Because of this, an inherent mix of patients exists in the prophylactic group, including those who were already declining from a nutritional standpoint, as well as patients who were doing well yet had certain risk factors for needing a PEG tube in the future. We attempted to avoid this heterogeneity in patient selection for PEG tubes by only addressing the patients who did not receive a prophylactic PEG tube. Unfortunately, this created its own bias by addressing only the healthier subpopulation of patients who attempted CRT without a PEG tube.
This study is limited by its retrospective nature, the long time frame (2004-2012), and the few reactively placed PEG tubes in the study population (11.7%). During this period, anatomical structures potentially influencing dysphagia were increasingly contoured and irradiation doses constrained, including the larynx and superior, middle, and inferior pharyngeal constrictors. As we understand more about irradiation effects on swallowing, symptomatic requirement for PEG tube placement may decrease further. We improved the heterogeneity in patient treatment by the inclusion of only patients who underwent CRT. However, by assessing solely the subset of patients who did not receive prophylactic PEG tubes (43.1%), our sample size also decreased (n = 128), which limited the strength of the statistical modeling, with only 15 total events.
In conclusion, among patients undergoing concurrent chemotherapy and intensity-modulated radiation therapy for oropharyngeal squamous cell carcinoma, a BMI of less than 25, accelerated irradiation fractionation, a tumor T classification of 3 or higher, and a cumulative cisplatin dose of 200 mg/m2 or higher led to an increased risk of reactively requiring the placement of a PEG tube during CRT. In the lower-risk subset of patients who did not require a PEG tube before treatment, the reactive placement of a PEG tube rapidly corrected their lost weight.
Submitted for Publication: February 25, 2013; final revision received July 12, 2013; accepted August 27, 2013.
Corresponding Author: Jimmy J. Caudell, MD, PhD, Department of Radiation Oncology, Moffitt Cancer Center, 12902 Magnolia Dr, Tampa, FL 33612 (email@example.com).
Published Online: October 17, 2013. doi:10.1001/jamaoto.2013.5193.
Author Contributions: Drs Strom and Caudell had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Strom, Trotti, Kish, McCaffrey, Caudell.
Acquisition of data: Strom, Trotti, Rao, Caudell.
Analysis and interpretation of data: Strom, Trotti, McCaffrey, Padhya, Lin, Fulp, Caudell.
Drafting of the manuscript: Strom, Trotti, Rao, Caudell.
Critical revision of the manuscript for important intellectual content: Strom, Trotti, Kish, McCaffrey, Padhya, Lin, Fulp, Caudell.
Statistical analysis: Strom, Trotti, Padhya, Lin, Fulp, Caudell.
Administrative, technical, and material support: Strom, Trotti, Caudell.
Study supervision: Trotti, Rao, McCaffrey, Padhya, Caudell.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This study was presented at the American Head and Neck Society 2013 Annual Meeting; April 11, 2013; Orlando, Florida.