The prevalence of symptomatic peripheral arterial disease in the adult population ranges between 0.6% and 9.2% and increases with age 12. Patients with peripheral arterial disease have an increased risk of cardiovascular morbidity and mortality, showing a similar risk factor profile to patients with other atherosclerotic diseases. In the non-pharmacological treatment of symptomatic peripheral arterial disease, peripheral vascular surgical interventions such as bypass grafting and endarterectomy play an important role 3. The long-term aim of surgical interventions is to prevent amputation of the limb and to reduce its resulting disability. According to current guidelines, surgical interventions are indicated for individuals with symptomatic disease (claudication), significant functional disability, resistance to exercise or pharmacotherapy, and a reasonable likelihood of symptomatic improvement 4. Whereas endarterectomy is an option in strictly localised disease, bypass grafting is generally used to circumvent severely stenosed sections of the peripheral arteries.

Different materials can be used for bypass grafting including autologous and homologous grafts from the saphenous vein or the human umbilical vein as well as prosthetic graft materials such as polytetrafluoroethylene (PTFE) or polyester (Dacron^®) grafts. Most studies so far have shown that autologous vein is superior to prosthetic graft materials in bypass surgery 567. A recent review comparing venous and PTFE bypass procedures reported 5-year primary patency rates of 74% and 39%, respectively 8. However, almost a third of patients eligible for peripheral bypass procedures do not have suitable veins, making the use of prosthetic materials necessary 9. Also, due to the high prevalence of cardiovascular co-morbidity, it may be required to keep suitable autologous veins for potential future use in coronary artery bypass grafting. The objective of our systematic review was, therefore, to identify available evidence and compare the effectiveness of the prosthetic bypass materials Dacron^® and PTFE in peripheral vascular bypass surgery and to perform meta-analyses, if possible.

We identified a total of 4421 publications in the search process. Excluding duplicates and non-relevant publications based on their title and/or abstract resulted in 419 publications for further quality assessment (Fig. 1). Of these, 9 publications were found to be relevant to the research question and fulfilled all inclusion criteria 121314151617181920.

All 9 RCT included in our analysis used primary patency as their main outcome measure. All studies provided a definition for the term patency in their methods section, with primary patency usually meaning unassisted/uninterrupted patency with no follow-up procedures on the bypass assessed by objective methods; one study, however, defined primary patency as assisted primary patency 12. Seven studies also reported results on secondary patency rates. Five studies present limb salvage or amputation rates, 4 perioperative mortality (<30 days), and 6 graft infection as further outcomes. Most of these endpoints were not defined in detail.

Table 1 gives an overview of study characteristics and main results of all RCT included in our analysis.

Of the 9 included studies, 7 studies showed no significant differences between Dacron^® and PTFE regarding primary patency, 1 study showed significantly higher patency rates for Dacron^® after 2 years 20, and 1 study showed significantly higher patency rates for PTFE after 2 years 17.

The 7 RCT reporting results on secondary patency yielded results similar to primary patency rates, e.g. no significant differences between the prosthetic materials were found in 6 studies, and 1 study showed significantly higher secondary patency after 2 years with Dacron^® grafts. Of the studies reporting results regarding limb salvage or amputation rates, 4 showed no significant difference, whereas 1 showed a significantly better rate for Dacron^® grafts. None of the studies found significant differences in terms of graft infection or perioperative mortality (<30 days).

The present meta-analysis indicated that there are no major differences in primary and secondary patency rates between the two prosthetic graft materials Dacron^® and PTFE. Of the 9 included studies, one study showed statistically significant higher patency rates for Dacron^® after 2 years 20, 1 study showed statistically significant higher patency rates for PTFE after 2 years 17, while 7 studies showed no statistically significant differences between the two materials regarding primary patency 12131415161819.

Our study complements a systematic Cochrane review on femoro-popliteal bypass surgery by Mamode and Scott comparing saphenous vein graft with PTFE or Dacron^®, human umbilical vein with PTFE, and PTFE vs. Dacron^® 21. However, only one study comparing PTFE with Dacron^® in above-knee popliteal grafting was identified 22. This study by Abbott et al. was published before the search period of our review. It did not show a significant difference regarding primary or secondary patency rates at the 3-year follow-up.

One of the studies in the present review compared gelatine-coated Dacron^® vs. collagen-coated Dacron^® showing no differences between the study groups 16. In another study heparin-bonded Dacron^® instead of bare Dacron^® graft material was used. The issue of different coatings might in itself affect patency outcomes of Dacron^® graft. However this question cannot adequately be assessed at the moment due to the lack of data.

Only 5 studies presented data in adequate life table format for performing a meta-analysis. Results of the meta-analysis might change, if more study data could be included, especially data of the trial by Jensen et al. 20 as this was the largest trial to date comparing PTFE vs. Dacron^®. However, it should be noted that patient recruitment and surgeries of that trial were performed up to 14 years before publication of the results (2 year follow up). For the meta-analysis no raw data with actual times of event or censoring could be used, as only aggregated life table data was available. This will overestimate event and censoring times in both groups, but should not have influenced the comparison of the two graft materials.

In the context of clinical routine care, physicians regularly face decisions as to which alternative treatment strategies to recommend and use. Their decisions should be guided by best possible evidence of previous studies. Hence, systematic reviews of good quality RCT have evolved as important tool of decision support. It is important to note, however, that the RCT included in our systematic review were limited by a number of methodological limitations, such as rather small sample sizes, different methods for determining patency rates, a lack of consideration paid to additional factors that might affect outcomes such as baseline differences between groups, and inadequate interpretations of non-significant results. Reporting of existing baseline differences between the groups in each trial was heterogeneous. Only 4 studies provided information about adjustment for baseline differences as potentially confounding factors 13141819. Group differences at baseline may thus have biased the results. In 4 trials adequate sample size calculations were reported 15181920. The other trials may have been too small to detect any differences in the graft materials. In addition, no trial used equivalence testing to show that both graft types were similarly effective. All studies described how primary patency was assessed and defined, and most studies used recommended objective standard methods 11. However, differences in patency rates between studies might have been affected by unequal assessments, while differences in patency between PTFE or Dacron^® grafts within studies should not be affected (assuming same assessment standards for patients receiving PTFE or Dacron^® grafts within the trial).

Sources of funding were not specified in 4 trials, with the remaining trials being funded by or having received grants from bypass graft manufacturers, which might introduce bias to the results 2324. No explicitly independent/non-manufacturer sponsored trials were found.

In this review only trials published in English or German were considered for inclusion. Even though we feel to cover the most important and qualitatively higher trials with this strategy, this might result in relevant articles and evidence published in other languages being ignored. Results might change if a body of evidence from manuscripts in other languages would be available. Since 4 of the 9 includes trials were conducted in non-English speaking countries (Austria, Germany, and Scandinavian countries), but were published in English, we assume this might not be the case.

Even in the light of methodological limitations of the included trials, the present meta-analysis and systematic review might offer a basis for clinical decision making in individual patients requiring peripheral vascular surgery. Between the two prosthetic materials PTFE and Dacron^® no clear advantage of one over the other could be seen. Further independently funded studies should address the issue of heparin-bonded grafts as well as the identification of subgroups of patients, in which there might be a benefit of one material. Studies should be sufficiently powered to be able to detect differences or equivalence of PTFE and Dacron^® grafts.

The authors declare that they have no competing interests.

SR and DE were responsible for the methodological design of the review, the literature selection and the quality assessment. SR carried out the data extraction, the summary of the findings, the meta-analysis and has participated in the writing of the manuscript. JMN has participated in the writing of the manuscript. TK and WG have revised the manuscript critically. HS advised the review as a clinical expert and has revised the manuscript critically. SW had the overall scientific responsibility and has participated in the writing of the manuscript. All authors read and approved the final manuscript.