Sandra Ottosson&Ulrika Lindblom&Peter Wahlberg&
Per Nilsson&Elisabeth Kjellén&Björn Zackrisson&
Eva Levring Jäghagen&Göran Laurell
Received: 13 September 2013 / Accepted: 9 March 2014 / Published online: 1 April 2014
# Springer-Verlag Berlin Heidelberg 2014
Abstract
Purpose Persistent severe swallowing dysfunction with aspi-ration is a common and sometimes overlooked sequelae after treatment for squamous cell carcinoma of the head and neck (SCCHN) and may impact food intake and nutritional status.
More knowledge is needed to increase the understanding of severe swallowing dysfunction as a risk factor for persistent nutritional deteriorations in SCCHN survivors. The purpose of the study was to investigate weight loss and body mass index (BMI) in relation to pharyngeal swallowing function in a long-term perspective in patients after SCCHN treatment.
Methods Data from 101 patients were available for the anal-yses. Swallowing function was assessed by videofluoroscopy at a mean of 71.6 months after the start of radiotherapy (RT).
Percent weight change (calculated with weight at the start of RT as the reference) and BMI at follow-up were the primary nutritional measures.
Results Aspiration was present in 48 of 101 patients (48 %).
Patients with aspiration had a significantly higher mean weight loss and a lower BMI (−10.9 % and 23.1, respectively) at follow-up compared with patients without aspiration (−2.8 % and 26.0, respectively). Patients with aspiration were unable to gain weight after 23 months. Only ten of 101 patients (10 %) were underweight at follow-up.
Conclusions Swallowing dysfunction with aspiration was re-lated to long-term weight loss and reduced BMI. Few patients were underweight despite the high prevalence of swallowing dysfunction.
Keywords Fluoroscopy . Swallowing dysfunction . Head and neck cancer . Weight loss . Body mass index
Introduction
Swallowing dysfunction is a common and sometimes overlooked severe complication after treatment for squamous cell carcinoma of the head and neck (SCCHN) [1]. The treatment strategies used in SCCHN are radiotherapy (RT), surgery, or a combination of both and an increasing use of anticancer drugs [2]. These treatment protocols have different impacts on tissues important for optimal swallowing function [1,3], and patients treated with combined treatment protocols are at high risk for developing a severe swallowing dysfunc-tion [4]. Other factors that are reported to impact swallowing function are tumor site, tumor stage, and RT protocol [4,5].
The reported prevalence of swallowing dysfunction after RT varies [1] and is believed to be an underreported and thereby sometimes untreated problem in patients with SCCHN [6]. It Eva Levring Jäghagen and Göran Laurell shared senior authorship
S. Ottosson (*)
Department of Clinical Sciences, Otorhinolaryngology, Umeå University, 901 87 Umeå, Sweden
e-mail: sandra.ottosson@kost.umu.se U. Lindblom:P. Wahlberg
Department of Clinical Sciences, Otorhinolaryngology, Lund University, Lund, Sweden
P. Nilsson:E. Kjellén
Department of Oncology and Radiation Physics, Skåne University Hospital, Lund University, Lund, Sweden
B. Zackrisson
Department of Radiation Sciences, Oncology, Umeå University, Umeå, Sweden
E. Levring Jäghagen
Department of Odontology, Oral and Maxillofacial Radiology, Umeå University, Umeå, Sweden
G. Laurell
Department of Surgical Sciences, Otolaryngology and Head & Neck Surgery, Uppsala University, Uppsala, Sweden
has also been shown that swallowing dysfunction can persist and sometimes even continue to deteriorate after the termina-tion of treatment [7,8].
Between the years of 1998 to 2006, a Swedish national, randomized, prospective, multicenter trial on patients with SCCHN was conducted (the ARTSCAN trial) [9,10]. The aim of the ARTSCAN trial was to evaluate the effect of conventional fractionation versus accelerated fractionation on loco-regional control, overall survival, and quality of life.
The 2-year follow-up did not show any significant differences in survival and loco-regional control between the two treat-ment arms [9]. The data from the present study are based on a follow-up of a cohort of patients in the ARTSCAN trial from two of the participating treatment centers.
Aspiration is a sign of severe swallowing dysfunction and a result of a variety of swallowing deficiencies. Aspiration can lead to pneumonia [11] and can be lethal [12]. Nutritionally, impaired swallowing function can result in restricted food intake and dependence on tube feeding [13,14]. Consequent-ly, malnutrition is a common problem in patients with SCCH N [15] and is significantly related to swallowing dysfunction caused by treatment [16]. However, the relation between long-term persistent swallowing dysfunction and nutritional param-eters has been poorly described in previous research. Thus, more knowledge is needed on severe swallowing dysfunction as a risk factor for the development of a poor nutritional status in SCCHN cancer survivors to increase the understanding of persistent nutritional deteriorations after SCCHN treatment.
Aim
The aim of the present study was to investigate the long-term impact of pharyngeal swallowing function on weight devel-opment and BMI in patients with head and neck cancer treated with RT.
Materials and methods
Patients in the ARTSCAN trial
Seven hundred and fifty patients (>18 years) were included in the ARTSCAN trial [10,9]. All tumors were M0 SCCHN of the oral cavity, oropharynx, hypopharynx, or larynx (except glottic T1 and T2) and classified according to the Union for International Cancer Control (UICC) TNM staging system [17]. Patients treated with chemotherapy within 3 months before the start of RT were excluded. The treatment regime was RT (conventional fractionation or accelerated fraction-ation) as a single modality treatment or as preoperative RT.
Conventional fractionation was given with 2.0 Gy/day for 7 weeks and accelerated fractionation was given with
2.0 Gy+1.1 Gy/day for 4.5 weeks. All patients received CT-based three-dimensional conformal RT and/or intensity-modulated RT, and both treatment protocols reached a total dose of 68 Gy. For more information about the ARTSCAN trial the reader is referred to the original manuscripts [10,9].
Patients from two participating treatment centers
Half of the patients (n=380) in the ARTSCAN trial were treated at two of the participating treatment centers involved in the present study. Fifteen months after the ARTSCAN trial was closed, surviving patients without recurrent disease who had been previously treated at the two centers were invited to participate in follow-up examinations of swallowing function.
At that time, 202 patients were available and 124 patients accepted participation (Fig.1).
ARTSCAN follow-up
In the ARTSCAN trial, acute and late treatment toxicities, treatment outcomes, and any additional treatments (re-irradi-ation, surgery, and/or anticancer drugs) were registered up to 5 years in surviving patients.
Ear, nose, and throat examination
The additional follow-up on 124 patients included a complete ear, nose, and throat (ENT) examination. The clinical exam-inations were performed by one ENT physician at each treat-ment center.
The oral and pharyngeal swallowing function was assessed using videofluoroscopic (VF) examinations [18]
(Fig. 2). Before examination, the patients answered routine clinical questions about symptoms of dysphagia. The ex-amination was performed in lateral and frontal projections that provided views of the oral cavity, the pharynx, and the upper esophageal sphincter. All patients sat in an up-right position comfortable for swallowing. The first stan-dard swallow consisted of a 7-mL liquid bolus of barium sulfate, and if the swallow was without severe aspiration, it was followed by a 10-mL liquid bolus and modified boluses of varying consistencies. Patients with severe symptoms of aspiration before examination initially swallowed iodine contrast media. All VF examinations were stored digitally or on videotape. The evaluations were performed by the same radiologist together with one of the two ENT physicians who performed the clin-ical examinations. The examinations were assessed for several parameters of swallowing dysfunction. In this study, we present results of severe swallowing dysfunction causing aspiration. Aspiration was defined as present when parts of the bolus penetrated into the larynx and passed below the vocal cords into trachea [18, 19] (Fig. 2).
Nutritional data
Information about weight at the start of RT and at 11 and 23 months after the termination of RT was registered in the ARTSCAN protocol and weight at follow-up was gathered at the time of the ENT examination. Weight at follow-up was used to calculate relative percent weight change with weight at the start of RT as the reference point. Height was collected
from the medical records and used for calculation of BMI (kg/
m2). The following cut-off values for BMI were used in the analyses: underweight BMI of <20 (BMI of <22 if≥70 years of age); normal weight BMI of 20–25 (BMI, 22–27 if
≥70 years of age); and overweight/obese BMI of >25 (BMI,
>27 if≥70 years of age) [20,21]. Patients with underweight and a previous weight loss of≥10 % at follow-up were defined as“at risk patients” for a poor nutritional status.
The nutritional regimes varied at the two participating centers during RT and the following convalescence period.
Nutritional support was given based on the physicians’ and dieticians’ evaluation and patient approval. All included pa-tients had oral intake without tube feeding at follow-up and information about previous use of tube feeding (nasogastric feeding tube, percutaneous endogastric gastrostomy, or both) was available from the study protocol and the ARTSCAN database.
Ethical approval
The study was approved by the Regional Ethical Review Board in Umeå, Sweden (EPN Umeå Dnr 07-023M) and all participants signed informed consent forms.
Statistical analyses
The Statistical Package for the Social Sciences (SPSS) version 21.0 was used for the statistical analyses. Univariate analyses were performed using the independent samples t test for comparison of continuous variables between two groups, one-way between-groups ANOVA for comparison of contin-uous variables between more than two groups, or the Fisher’s exact test for comparison between groups. Variables found to be significant in the univariate analyses were used as contin-uous or dichotomized variables in a multivariate model using linear regression analysis. The unstandardized regression co-efficients (B) represent an increase (positive values) or a decrease (negative values) in weight (percentage points) or BMI. Repeated measures ANOVA was used to analyze the change in mean weight over time. All tests were two-sided and a p value less than 0.05 was considered significant.
Results
Eligibility and patient characteristics
Nine patients were excluded due to; ischemic stroke (n=2), additional tumor disease (n=2), tube feeding dependence (n=
3), or laryngectomy (n=2). Three patients died between the ENT and the VF examinations and an additional 11 patients were not examined due to practical reasons or had missing weight data registrations at follow-up. Therefore, 101 of the Fig. 1 Patient eligibility within the studied cohort
Fig. 2 Aspiration. Lateral videofluoroscopic view of the pharynx, lar-ynx, and the upper part of the trachea after a swallow. The aspirated bolus (AB) has penetrated into the larynx (L) and continued below the vocal cords (VC) into the trachea (T). Epiglottis (E)
original 124 patients (82 %) had weight data and BMI regis-trations at the start of RT and at the follow-up and were defined as the study cohort 1 (Fig.1).
Seventy-six of the 101 patients (75 %) were men and the median age for the cohort was 62 years (range, 34–84 years) (Table1). Tumor of the oropharynx was the most common site (n=62, 61 %) and 74 patients (73 %) had stage III or IV tumors. Fifty-eight patients (57 %) were treated with acceler-ated fractionation, 43 patients (43 %) were treacceler-ated with con-ventional fractionation, and 58 patients (57 %) underwent surgery after preoperative RT. In total, 59 patients (58 %) received tube feeding during treatment and/or the convales-cence period and two were given tube feeding at the start of RT.
Swallowing function
The ENT examination took place at a mean of 69.3 months (±29.6) after the start of RT and the VF was performed at a
mean of 71.6 months (±28.3) after the start of RT. Aspiration was diagnosed with VF in 48 patients (48 %) at follow-up and 33 of these patients were diagnosed with silent aspiration, as aspiration did not elicit a cough. Five of 48 patients had been treated for pneumonia. No significant difference was found in previous tube feeding use between the non-aspiration group and the aspiration group (p=0.545).
Impact factors for long-term weight loss and body mass index Weight loss
The mean weight loss between the start of RT and follow-up was 6.6 % (±10.5 %). The weight loss was analyzed in univariate analyses together with patient, tumor, treatment, and nutritional factors (Table1). The following factors were found to be significantly related to weight loss: gender, tumor site, surgery, BMI at the start of RT, and posttreatment aspi-ration. A multiple linear regression was used to make a model
Table 1 Characteristics for 101 patients (n (%)) and different predictive factors for relative percent weight change (mean (SD)) between the start of radiotherapy (RT) and follow-up (mean, 69.3 months) and body mass index (BMI) at follow-up (mean (SD))
Patient characteristic n (%) Weight change (%) p value* BMI p value*
Mean (SD) Mean (SD)
Age <65 years 60 (59.4) −5.4 (10.6) 0.154 25.1 (4.4) 0.125
≥65 years 41 (40.6) −8.4 (10.2) 23.9 (3.3)
Gender Male 76 (75.2) −7.8 (10.0) 0.043 24.5 (3.6) 0.563
Female 25 (24.8) −2.9 (11.5) 25.0 (5.3)
Tumor site Oropharynx 62 (61.4) −9.9 (9.3) <0.001 24.5 (4.2) 0.862
Oral cavity 20 (19.8) −5.3 (6.4) 24.8 (2.9)
Larynx 11 (10.9) −1.0 (13.3) 25.4 (5.1)
Hypopharynx 8 (7.9) 7.3 (10.0) 24.0 (4.3)
Tumor stage I 11 (10.9) −1.4 (9.2) 0.103 25.5 (3.3) 0.182
II 16 (15.8) −3.4 (11.9) 26.3 (5.8)
III 28 (27.7) −7.0 (9.2) 23.7 (3.6)
IV 46 (45.5) −8.8 (10.7) 24.4 (3.6)
Fractionation schedule Accelerated fractionation 58 (57.4) −6.2 (10.7) 0.646 24.2 (3.7) 0.193
Conventional fractionation 43 (42.6) −7.2 (10.4) 25.2 (4.5)
Surgery No surgery No 43 (42.6) 43 (42.6) −3.4 (9.7) 0.007 25.3 (4.9) 0.145
Resection of primary tumor Yes 11 (10.9) 58 (57.4) −9.1 (10.6) 24.1 (3.2)
Neck dissection 42 (41.6)
Resection of primary tumor and neck dissection
5 (5.0)
Aspiration No 53 (52.5) −2.8 (8.1) <0.001 26.0 (4.3) <0.001
Yes 48 (47.5) −10.9 (11.3) 23.1 (3.2)
BMI start RTa Underweight 5 (5.0) −0.2 (6.3) 0.020 19.0 (1.8) 0.025**
<0.001***
Normal weight 29 (28.7) −4.3 (8.2) 22.7 (1.9)
Overweight/obesity 61 (60.4) −9.4 (10.6) 26.3 (4.0)
Earlier tube feeding use No 42 (41.6) −6.8 (8.9) 0.895 25.1 (3.8) 0.312
Yes 59 (58.4) −6.5 (11.6) 24.3 (4.2)
*p<0.05 was considered significant in the independent samples t test or one-way ANOVA; **p value when comparing underweight against normal weight; ***p value when comparing underweight against overweight/obesity or normal weight against overweight/obesity
aUnderweight, BMI<20 (BMI<22 if≥70 years of age); normal weight, BMI=20–25 (BMI=22–27 if ≥70 years of age); overweight/obese, BMI>25 (BMI>27 if≥70 years of age)
of the significant variables from the univariate analyses to predict weight loss. Percent weight loss was used as the dependent variable (numerical) and the independent variables were gender (male/female), surgery (yes/no), tumor site (oro-pharynx, oral cavity, larynx, or hypopharynx), BMI at the start of RT (underweight, normal weight, or overweight/obesity), and posttreatment aspiration (yes/no). The coefficient of de-termination R2for the model was 0.462 (p<0.001). Aspira-tion, BMI at the start of RT, and tumor site were significantly predictive for weight loss. Patients without aspiration had significantly less weight loss than patients with aspiration (B=−7.278; p<0.001). Patients with overweight or obesity according to the BMI classification demonstrated significantly greater weight loss than patients with normal weight (B=
4.419; p=0.015) or underweight (B=8.950; p=0.010). More-over, patients with tumor of the oropharynx had a significantly greater weight loss compared with patients having tumors of the larynx (B=6.921; p=0.017) or hypopharynx (B=13.399;
p<0.001). No significant difference was seen between pa-tients with tumors of the oropharynx and oral cavity (B=
2.993; p=0.180).
Body mass index
Mean BMI was 26.5 (±4.3) and 24.6 (±4.0) at the start of RT and at follow-up, respectively. Seven patients (7 %) were underweight, 33 patients (33 %) were normal weight, and 61 patients (60 %) were overweight/obese at the start of RT. No significant difference was seen between patients with and without aspiration according to the different BMI groups at the start of RT (p=0.858). At follow-up, 10 patients (10 %) were underweight, 54 patients (53 %) were normal weight, and 37 patients (37 %) were overweight/obese. BMI at follow-up was analyzed in univariate analyses together with patient, tumor, treatment, and nutritional factors (Table1). Posttreat-ment aspiration and a low BMI at the start of RT were found to be significantly related to a low BMI at follow-up. A multiple linear regression was used to make a model of the significant
variables from the univariate analyses to predict BMI. BMI was used as the dependent variable (numerical) and the inde-pendent variables were posttreatment aspiration (yes/no) and BMI at the start of RT (underweight, normal weight, or overweight/obesity). The coefficient of determination R2for the model was 0.421 (p<0.001). Both variables included in the model were significantly predictive for BMI at follow-up.
Patients without aspiration had a significantly higher BMI at follow-up than patients with aspiration (B=−2.620; p<0.001).
Moreover, patients with overweight or obesity according to the BMI classification demonstrated significantly higher BMI at follow-up than patients with normal weight (B=−3.517, p<0.001) or underweight (B=−7.019; p<0.001).
The patients were divided into four groups according to weight loss (weight stability or gain, weight loss of 2–10 %, weight loss of 10–20 %, and weight loss of ≥20 %) (Fig.3).
Ten patients (10 %) with underweight were evenly distributed between the four weight groups (weight stability or gain, n=4;
weight loss of 2–10 %, n=2; weight loss of 10–20 %, n=2;
and weight loss of≥20 %, n=2). Seven of these patients had aspiration and four patients demonstrated a previous weight loss of≥10 %. None of the patients in the non-aspiration group had lost≥20 % of pretreatment weight at follow-up compared with eight patients in the aspiration group.
Change in weight over time
In order to evaluate change of weight over time in relation to pharyngeal swallowing function a subsample of patients with subsequent data were included. Forty-nine patients (49 %) had weight data registrations at four different time points: the start of RT, 11 and 23 months after the termination of RT and the final follow-up (69.3 months) and were defined as the study cohort 2 (Fig.1). Mean weight (kg) changed significantly over time (p<0.001) in cohort 2. There was a significant decrease or increase in mean weight (kg) between each of the different time points (p<0.001) except for mean weight change be-tween 23 months and the final follow-up (p=1.000). The
Fig. 3 Relative percent weight change from the start of radiotherapy (RT) to follow-up (mean, 69.3 months) stratified into four weight groups and combined with information on body mass index (BMI) in patients with (n=48) and without (n=53) aspiration
percent weight change over time was recorded for patients with and without aspiration, and both groups had a rapid weight loss during and after treatment (Fig.4). At 11 months after the termination of RT, the weight started to increase for both groups but did not return to pretreatment values at the final follow-up. In contrast to patients without aspiration that continued to regain weight between 23 months and the final follow-up (2.6 %), patients with aspiration lost weight (−2.7 %; p=0.006).
Discussion
The present cohort study highlights that silent aspiration is common in long-term SCCHN survivors. The main finding was that aspiration is an important impact factor for both decreased BMI and weight several years after the termination of treatment for patients with head and neck cancer. Forty-eight patients (48 %) aspirated during VF and this is some-what higher than the prevalence found in earlier studies (22–
44 %) with similar follow-up times [7, 22]. Patients with aspiration had a significantly higher mean weight loss and a lower BMI (−10.9 % and 23.1, respectively) compared with patients without aspiration (−2.8 % and 26.0, respectively).
All patients in the present cohort were considered tumor-free.
The European Society for Clinical Nutrition and Metabo-lism (ESPEN) guidelines for nutritional screening point out weight loss and decreased BMI as important tools that should be used in clinical practice to find patients at risk of develop-ing a poor nutritional status [20]. No generally accepted guidelines are available on how critical weight loss should be defined in the long term, but weight loss of >5 % during 1 to 3 months or >10 % after 6 months is often used in the short term to identify patients that run a high risk of developing a poor nutritional status [20, 23]. Unintentional weight loss
during disease is mainly loss of fat free mass [24] and invol-untary weight loss in SCCHN survivors should therefore be the subject for thorough investigation and actions. However, one apparent finding in the present cohort was that before treatment and at the long-term follow-up 60 and 37 % of the patients, respectively, were defined as overweight or obese.
Only ten patients (10 %) were underweight at follow-up. Four of the patients with underweight displayed a previous weight loss of≥10 % and are therefore suggested to be defined as “at-risk patients” for a poor nutritional status. These findings allow us to state that patients with aspiration displayed a lower BMI and a higher weight loss than patients without aspiration but“at risk patients” for a poor nutritional status were still rare in the present cohort. However, to be able to elaborate on this statement appropriately, correct assessment of the patients nutritional status using, for example, the Patient-Generated Subjective Global Assessment (PG-SGA) [25] and informa-tion on when and for how long the weight loss was displayed,
Only ten patients (10 %) were underweight at follow-up. Four of the patients with underweight displayed a previous weight loss of≥10 % and are therefore suggested to be defined as “at-risk patients” for a poor nutritional status. These findings allow us to state that patients with aspiration displayed a lower BMI and a higher weight loss than patients without aspiration but“at risk patients” for a poor nutritional status were still rare in the present cohort. However, to be able to elaborate on this statement appropriately, correct assessment of the patients nutritional status using, for example, the Patient-Generated Subjective Global Assessment (PG-SGA) [25] and informa-tion on when and for how long the weight loss was displayed,