Introduction
Surgical resection of laryngeal malignancies and reconstruction of the subsequent defects in the aerodigestive tract are among the most challenging of oncological procedures; they are associated with substantial morbidity and mortality.
Pharyngocutaneous fistula formation, the most common complication following laryngectomy, has adverse effects on post-operative rehabilitation, adjuvant therapy and overall survival.Reference Markou, Vlachtsis, Nikolaou, Petridis, Kouloulas and Daniilidis1 Neopharyngeal stricture, another common complication, negatively impacts patients’ quality of life by impeding speech and swallow.Reference Lavertu, Guay, Meeker, Kmiecik, Secic and Wanamaker2
Several surgical strategies have been proposed to reduce the incidence of these complications. This systematic review summarises the evidence available for the routine use of salivary bypass tubes and the effect on fistula and stricture rates.
Materials and methods
Search strategy
A systematic search was conducted, on the 5th July 2016, of the Embase, Medline, and Cochrane Library databases, from 1946 – current. The following search terms were used: ‘salivary bypass tube’, ‘salivary tube’, ‘salivary stent’, ‘Montgomery tube’ and ‘Montgomery stent’. Only articles written in the English language and comprising human subjects were included. The reference lists of relevant articles were also searched.
Study selection
Included were all peer-reviewed published studies in which salivary bypass tubes were placed routinely during laryngectomy. No randomised controlled trials were found. Individual case reports, letters and conference abstracts were excluded.
Study evaluation
The included studies were graded according to the Oxford Centre for Evidence-based Medicine scheme. They were then assessed independently by the first two authors for risk of bias, according to the Cochrane Handbook for Systematic Reviews of Interventions.Reference Higgins and Green3 Discrepancies were referred to the senior author.
Results
The systematic search yielded 99 studies. Ninety of these studies were excluded; 30 were duplicates, and 60 were case reports, conference abstracts or letters (Figure 1). The included studies consisted of six case series (level 4 evidence) and three retrospective case–control studies (level 3b evidence). In total, 383 patients were investigated, of whom 204 received a salivary bypass tube.
Fig. 1 Results of the systematic search.
Within most of the studies, there was heterogeneity of type and extent of surgical resection; however, the majority of patients had either total or partial pharyngectomy, in addition to total laryngectomy.
The findings of the authors’ risk of bias assessment for each included study are shown in Table I.Reference Punthakee, Zaghi, Nabili, Knott and Blackwell4–Reference Jegoux, Ferron, Malard and Espitalier12 Overall, the included studies had a moderate-to-high risk of bias, because of the retrospective design and the heterogeneity of study populations. Confounding factors included prior chemoradiotherapy, use of flap versus primary closure, flap type, and primary versus secondary surgery.
Table I Authors’ risk of bias assessment for each included study*
* Assessment of the key domains, as set out in the Cochrane Handbook for Systematic Reviews of Interventions.Reference Higgins and Green3
Fistula rate
Of the three retrospective case–control studies, Bondi et al.Reference Bondi, Giordano, Limardo and Bussi5 (n = 53) reported a significantly lower fistula rate with salivary bypass tube use (45 per cent vs 9 per cent). Of note, the salivary bypass tube group had a higher proportion of flap repairs (i.e. inherently higher fistula risk). Punthakee et al.Reference Punthakee, Zaghi, Nabili, Knott and Blackwell4 (n = 103) found a significant reduction in fistula rates with salivary bypass tube use on univariate analysis; however, the effect was not significant once multivariate analysis was used to account for potential confounding factors. The study was substantially underpowered to detect the treatment effect. León et al.Reference León, Quer and Burgués6 (n = 61) reported a non-significant reduction in fistula rate with salivary bypass tube use. The results are summarised in Table II.Reference Punthakee, Zaghi, Nabili, Knott and Blackwell4–Reference León, Quer and Burgués6
Table II Summary of case–control study* results
* Level 3b evidence. †On univariate analysis. ‡On multivariate analysis, performed to account for potential confounding factors. SBT = salivary bypass tube; PMMF = pectoralis major myocutaneous flap
Data from the six case series were categorised according to the type of flap used; López et al.Reference López, Obeso, Camporro, Fueyo, Suárez and Llorente10 utilised two flap types. Two case series,Reference Varvares, Cheney, Gliklich, Boyd, Goldsmith and Lazor7, Reference López, Obeso, Camporro, Fueyo, Suárez and Llorente10 comprising 44 patients, used a radial free forearm flap and salivary bypass tube; 18 per cent developed fistulae. Two case seriesReference Murray, Gilbert, Vesely, Novak, Zaitlin-Gencher and Clark9, Reference López, Obeso, Camporro, Fueyo, Suárez and Llorente10 (total n = 45) used an anterolateral thigh flap and salivary bypass tube; 2 per cent developed fistulae. Three case seriesReference Spriano, Pellini and Roselli8, Reference Fabian11, Reference Jegoux, Ferron, Malard and Espitalier12 (total n = 77) used a pectoralis major myocutaneous flap and salivary bypass tube; a total of 15.6 per cent developed fistulae. These results are summarised in Table III, and compared with quoted rates for fistulae in the wider literature.Reference Varvares, Cheney, Gliklich, Boyd, Goldsmith and Lazor7–Reference Guimarães, Aires, Dedivitis, Kulcsar, Ramos and Cernea18 Study heterogeneity prevented meaningful meta-analysis.
Table III Fistula rates in case series* using salivary bypass tubes versus rates in wider literature without salivary bypass tubes, by type of flap
* Level 4 evidence. SBT = salivary bypass tube; RFFF = radial free forearm flap; ALT = anterolateral thigh flap; PMMF = pectoralis major myocutaneous flap
Stricture rate
Only one case–control study, by Punthakee et al.,Reference Punthakee, Zaghi, Nabili, Knott and Blackwell4 assessed the effect of salivary bypass tube use on neopharyngeal stricture rates; they found no association. Two case seriesReference Varvares, Cheney, Gliklich, Boyd, Goldsmith and Lazor7, Reference López, Obeso, Camporro, Fueyo, Suárez and Llorente10 (n = 44) reported an average stricture rate of 9 per cent using a radial free forearm flap and salivary bypass tube. Two case seriesReference Murray, Gilbert, Vesely, Novak, Zaitlin-Gencher and Clark9, Reference López, Obeso, Camporro, Fueyo, Suárez and Llorente10 (n = 45) reported an average stricture rate of 6.6 per cent using an anterolateral thigh flap and salivary bypass tube. Three case seriesReference Spriano, Pellini and Roselli8, Reference Fabian11, Reference Jegoux, Ferron, Malard and Espitalier12 (n = 77) reported an average stricture rate of 14 per cent using a pectoralis major myocutaneous flap and salivary bypass tube. Table IV summarises these results.Reference Varvares, Cheney, Gliklich, Boyd, Goldsmith and Lazor7–Reference Azizzadeh, Yafai, Rawnsley, Abemayor, Sercarz and Calcaterra13, Reference Clark, Gilbert, Irish, Brown, Neligan and Gullane15, Reference Yu, Hanasono, Skoracki, Baumann, Lewin and Weber16, Reference Nakatsuka, Harii, Asato, Ebihara, Yoshizumi and Saikawa19, Reference Scharpf and Esclamado20
Table IV Stricture rates in case series* using salivary bypass tubes versus rates in wider literature without salivary bypass tubes, by type of flap
* Level 4 evidence. SBT = salivary bypass tube; RFFF = radial free forearm flap; ALT = anterolateral thigh flap; PMMF = pectoralis major myocutaneous flap
Safety
Of the 204 patients covered in this review, who received a salivary bypass tube during laryngectomy, there were 2 arterial bleeds, 4 distal migrations and 7 proximal migrations. None of these events were fatal; the arterial bleed rates were similar between patients with and without a salivary bypass tube. As a result of the stent migrations, several authors reported securing the salivary bypass tube to a nasogastric tube sutured at the nasal septum.
Three case reports of salivary bypass tube related adverse events exist, all fatal. One arterioesophageal fistula occurred following laryngectomy due to a retroesophageal subclavian artery.Reference Inman, Kim and McHugh21 A distal migration of salivary bypass tube caused intestinal perforation following laryngectomy.Reference Bitter, Pantel, Dittmar, Guntinas-Lichius and Wittekindt22 An aortoesophageal fistula occurred in a paediatric patient following the long-term use of a salivary bypass tube to reconstruct oesophageal atresia.Reference McWhorter, Dunn and Teitell23
Discussion
Summary of main results
The largest case–control studyReference Punthakee, Zaghi, Nabili, Knott and Blackwell4 (n = 103) showed a reduction in fistula rates with salivary bypass tube use. This finding was significant on univariate analysis, but not on multivariate analysis that attempted to correct for confounding factors, such as flap type and prior chemoradiotherapy. This could indicate a true lack of an effect for salivary bypass tube treatment or reflect the underpowered nature of the study (188 participants were required to achieve 80 per cent power). All patients in this study had flap reconstruction, and thus were, arguably, at greater risk of fistulae. A second case–control studyReference León, Quer and Burgués6 found no significant effect; this study reported a high incidence of fistulae overall in both groups (54 per cent). The third case–control studyReference Bondi, Giordano, Limardo and Bussi5 showed a significant reduction in the fistula rate with salivary bypass tube use. The salivary bypass tube group had higher rates of pre-operative radiotherapy and higher rates of flap repair. It is important to note that in this study, the authors selected patients they considered at high risk for fistula development (based on the extent of the tumour and prior radiotherapy). This may explain the large treatment effect observed with salivary bypass tube use in that cohort.
Two case seriesReference Varvares, Cheney, Gliklich, Boyd, Goldsmith and Lazor7, Reference López, Obeso, Camporro, Fueyo, Suárez and Llorente10 (total n = 44), focusing on salivary bypass tube use with a radial free forearm flap, reported an average fistula rate of 18 per cent, as compared to quoted rates in the literature of 20–53 per cent.Reference Azizzadeh, Yafai, Rawnsley, Abemayor, Sercarz and Calcaterra13–Reference Clark, Gilbert, Irish, Brown, Neligan and Gullane15 Two case series,Reference Murray, Gilbert, Vesely, Novak, Zaitlin-Gencher and Clark9, Reference López, Obeso, Camporro, Fueyo, Suárez and Llorente10 investigating salivary bypass tube use with an anterolateral thigh flap, reported that 2 per cent of 45 patients developed fistulae, compared to quoted rates of 9–30 per centReference Clark, Gilbert, Irish, Brown, Neligan and Gullane15–Reference Morrissey, O'Connell, Garg, Seikaly and Harris17 for anterolateral thigh flaps without a salivary bypass tube. This could suggest some treatment benefit. However, the heterogeneity of the patients included in these studies and in the wider literature prevents any meaningful comparison via meta-analysis. The average rate of fistula found with a pectoralis major myocutaneous flap and salivary bypass tube in three studiesReference Spriano, Pellini and Roselli8, Reference Fabian11, Reference Jegoux, Ferron, Malard and Espitalier12 was 15.6 per cent, similar to that reported in a large meta-analysis of a pectoralis major myocutaneous flap series (19.4 per cent).Reference Guimarães, Aires, Dedivitis, Kulcsar, Ramos and Cernea18
With regard to the effect of salivary bypass tube use on stricture formation, the largest and most robust study included in this review, by Punthakee et al.,Reference Punthakee, Zaghi, Nabili, Knott and Blackwell4 found no association. Seven case series reported stricture rates with salivary bypass tube use ranging from 3 to 16 per cent. In comparison, the rates in the literature are: 20–36 per cent for a radial free forearm flap,Reference Azizzadeh, Yafai, Rawnsley, Abemayor, Sercarz and Calcaterra13, Reference Nakatsuka, Harii, Asato, Ebihara, Yoshizumi and Saikawa19, Reference Scharpf and Esclamado20 6–24 per cent for an anterolateral thigh flapReference Clark, Gilbert, Irish, Brown, Neligan and Gullane15, Reference Yu, Hanasono, Skoracki, Baumann, Lewin and Weber16 and 12 per cent for a pectoralis major myocutaneous flap.Reference Yu, Hanasono, Skoracki, Baumann, Lewin and Weber16 The wide range of stricture rates without salivary bypass tube use precludes direct statistical comparison, but the rates seem broadly similar with or without salivary bypass tube use.
Level of evidence
The level of evidence found for this review was at best level 3b (three studies) and otherwise level 4 (six studies). The included studies are all of relatively weak design, and at high risk of bias, because of factors such as: retrospective data collection; lack of blinding of surgeons, patients or data analysts; and small patient groups with multiple confounding factors. In particular, cases and controls were non-contemporaneous, with the salivary bypass tube patients being operated on later than the non-salivary bypass tube patients, such that the salivary bypass tube groups could have benefited from technical and surgical advances.
Quality and completeness of evidence
The heterogeneity of the patient populations within each study significantly limits the ability to draw any firm conclusions regarding the effect of salivary bypass tubes. In particular, the variation within and between studies in regard to the extent of surgical resection limits direct comparison. Conversely, these relatively small and diverse patient cohorts could be argued to reflect the reality of advanced laryngeal cancer populations, and as such provide practical and relevant information.
Strategies to reduce the risk of bias in this review included the use of a robust search strategy, in conjunction with a medical librarian, which was then run twice to ensure reproducible results. However, because of practical constraints, the search was limited to articles written in the English language. Two authors independently assessed the included studies for risk of bias according to the criteria specified in the Cochrane Handbook for Systematic Reviews of Interventions. However, as the included studies themselves had a high risk of bias because of their design, this review will inherently carry a risk of bias. To our knowledge, there are no other reviews on this topic with which to compare.
Implications for practice and research
The data in this review are not sufficiently robust to support recommendations in clinical practice. However, the findings suggest that salivary bypass tube use in laryngectomy might benefit certain patients who are at high risk of fistula formation, and this may inform surgical decision making in individual challenging cases. A large, multicentre cohort or, ideally, randomised controlled trial, is needed. This is warranted by the existing evidence, to examine salivary bypass tube use in patients stratified into different risk categories.
