Venous leg ulcers

Cox regression: therapy HR 2.4 (1.19 to 4.71), P < .01 for preparation; HR 3.22 (1.66 to 6.21), P < .000 for healing.

Costs, in United States dollars: V.A.C.-related products $847; bandages and dressings $2,391 against $4,770; personnel $583 against $508; nursing $124 against $175; totals printed as $3,881 against $5,452, P = .001.

Adverse events by type: cutaneous damage secondary to therapy 7 against 2; pain 3 against 1; erysipelas 1 against 0; wound infection 0 against 1; donor-site bleeding 0 against 2; non-healing ulcer 1 against 1.

CAT: OCEBM 2 · Start High (randomised trial with computer-generated sequence, permuted blocks, stratification and central telephone allocation). -1 risk of bias: Masking was not possible, and the two endpoints that matter most here - time to complete healing and, critically, wound bed preparation - are clinician judgements made by assessors who could see the device. The trial mitigates this as far as it can, with 'the same independent research physician and consultant dermatologist twice a week', but it cannot eliminate it. The primary analysis is also not what it claims to be: the abstract states analysis was 'on the intention-to-treat basis', while the primary analysis used 28 and 26 of 60 randomised patients, and 23 and 24 at twelve months. -1 imprecision: 60 patients, powered on retrospective data to detect a 7-day difference in preparation time. Not downgraded for indirectness despite the population being hospitalised until healing, because the wound biology transfers even where the care model does not. Net low. The endpoint this library came for is wound bed preparation, and its definition is the most useful sentence in the paper: preparation ran from surgical debridement to punch grafting, and 'A wound was considered to be prepared when granulation tissue covered 100% of the surface and wound secretion was minimal' - the Falanga criteria. Median 7 days (95% CI 5.7 to 8.3) against 17 days (10 to 24), p = .005, with 90% of V.A.C. ulcers prepared within 14 days against 37% of controls. This library has established that no study anywhere compares one NPWT stopping criterion against another. This trial does not test one either - but it is the only paper this library holds that defines one and measures the time to reach it. It should be cited as a defensible target for when to stop, and never as a bias-free effect estimate, because it was a secondary endpoint assessed unblinded.

Figures: Figures checked

One trial. Sixty participants. Every outcome low certainty.

The single included study is Vuerstaek 2006, named in the flow-diagram narrative and in the characteristics-of-included-studies table, and already held in this library as Vuerstaek 2006. There is therefore no independent corroboration in this review of anything: the Cochrane review and the Vuerstaek page describe the same 60 patients.

Median time to healing is given as 28 days (95% CI 25.5 to 32.5) against 45 days (36.2 to 53.8). Median time to preparation for surgery, 7 days (95% CI 5.7 to 8.3) against 17 days (10 to 24).

The proportion-healed figure is the review's own calculation, not the trial's. Six participants were lost to follow-up; applying its pre-stated rule that a randomised participant absent from an analysis is counted as unhealed, the review arrives at 29/30 in both arms and RR 1.00. Both GRADE downgrades on that row are for imprecision and for indirectness of the care pathway.

CAT: OCEBM 1 · Start High as a Cochrane review of randomised evidence, then apply the review's own GRADE table, which this appraisal adopts. −1 IMPRECISION: the entire evidence base is 60 participants in a single trial; the healing-proportion interval, RR 1.00 (0.91 to 1.10), spans a 10% reduction to a 10% increase, and the hazard ratio for time to healing, 3.2 (1.7 to 6.2), is a fourfold-wide interval. −1 INDIRECTNESS, and this is the load-bearing downgrade: Every participant in both arms received a punch skin-graft transplantation, and every participant remained a hospital in-patient until the wound healed. The review's own footnote reads that 'the applicability of the care pathway (that is use of skin punch grafts and treatment of patients as inpatients until healed) was considered limited', and the plain-language summary states that 'The study's results are not relevant for leg ulcers that are not being prepared for skin grafts.' Net Low on every outcome. The sentence that matters most in this document is the one that reports an absence: 'There is no RCT evidence on the effectiveness of NPWT as a primary treatment for leg ulcers.' Not downgraded further for inconsistency OR publication bias, because with one trial neither can be assessed — which is itself a reason for caution rather than reassurance, and the authors cite Peinemann 2008 on unpublished NPWT data. The single included trial is Vuerstaek 2006, already held in this library, and it was funded by the Dutch department of the device manufacturer. The review and the trial page therefore describe the same 60 patients; there is no independent corroboration anywhere in this review.

Figures: Figures checked

Section 6.2.1 adds that an ABI > 0.8 "may be considered as normal and allows commencement of full compression therapy", and that in diabetics with incompressible arteries arterial duplex or toe pressure may be required. Section 6.3.2 gives the physiological basis: inelastic multilayer bandages or adjustable garments "exerting a high pressure ≥ 40 mmHg" should be applied; the two-stocking kit exerts about 40 mmHg supine rising to almost 50 mmHg standing; the static stiffness index of inelastic systems exceeds 10 mmHg.

CAT: Clinical practice guideline — AGREE II applied; GRADE not applicable to the document as a whole. Recommendation 70 (≥40 mmHg) carries ESC Class I, Level A on three cited sources; Recommendation 72 (no sustained compression below ankle pressure 60 mmHg / toe 30 mmHg / ABI 0.6) is Class III, Level C consensus · A guideline is not graded for effect; its recommendations inherit the certainty of what they cite, and the one this library needs is well founded. Recommendation 70 — at least 40 mmHg at the ankle by multilayer or inelastic bandage or adjustable garment — is Class I Level A on the O'Meara 2012 Cochrane review of compression, the Dolibog 2014 RCT and the Mosti 2020 RCT; the underlying Cochrane evidence for compression versus none is consistent and large, and the pressure target rests on haemodynamic measurement (static stiffness index, standing versus supine pressure) as well as on trial arms delivering that pressure. The ESC class/level system is not GRADE: Level A means 'multiple RCTs or meta-analyses' regardless of their risk of bias, so a Level A here is a statement about design and volume, not about certainty after downgrading; this library carries the recommendation as strong and its evidence base as moderate on its own reading. Recommendation 72, the lower limit below which sustained compression is not to be used (ankle pressure < 60 mmHg, toe < 30 mmHg, ABI < 0.6), is consensus, Class III Level C — a safety threshold by expert agreement, which is the usual and defensible basis for a harm threshold. The NPWT statement in section 6.2.3.3 is not a numbered recommendation; it is narrative resting on one Cochrane review (Dumville 2015, held as Dumville 2015) and it says what that review says: no RCT of NPWT as primary VLU treatment, one trial of NPWT as bed preparation before punch grafting with limited applicability. That is consistent with this library's guarded status for the venous ulcer. AGREE II domain scores are recorded in the amstar2 field; the weakest domains are applicability (no cost or resource statements by the authors' own design) and editorial independence (disclosures filed, not published).

Figures: Figures checked

CAT: OCEBM 5 · This is a narrative review and there is no pooled effect to grade. What can be graded is the ALGORITHM, node by node, and the result is that its most usable content carries no evidence at all. The numeric selection thresholds are uncited: the 12 cm-squared floor below which NPWT should not be used, the 40 cm-squared threshold at which benefit 'quite frankly' begins (which the authors themselves call 'somewhat arbitrary'), the composite treatment trigger of under 40% reduction at 4 weeks OR duration over 1 year OR over 40 cm-squared (asserted as 'adequate data exist' with no REFERENCE, and no cited trial enrolled by these criteria), and the 100 cm-squared threshold for instillation. The compression statements are also uncited - 'The primary treatment of VLU remains appropriate compression' and 'increasing venous return with multilayer compression is mandatory for good ulcer healing' both carry no citation, as does 'most clinicians who treat VLUs with adjunctive NPWT use it in conjunction with multilayer compression', for which no survey exists. Also uncited: that a nonstick contact layer is 'paramount'; that better outcomes follow more aggressive haemostatic debridement ('Presumably ... It seems that'); the tumescent lidocaine formula of 440 mL saline plus 60 mL 1% lidocaine plus 24 mL bicarbonate, which is the authors' own practice with no dose-safety citation; and that many single-use devices have a 2-week battery life, which drives the recommended 2-week treatment block. The one result that will bear weight is Vuerstaek 2006 - a randomised trial of 60 hospitalised patients in which NPWT halved wound-bed preparation time from 17 to 7 days (p = .005) and shortened time to complete healing from 45 to 29 days (p = .0001). Everything else on bed preparation and bolstering in venous disease rests on case series of 9 to 15 patients, several from the senior author's own unit.

Figures: Figures checked