Reviews/Novel Antibiotics/Lipoglycopeptides reduce length of stay
Lipoglycopeptides · NOVEL-AB-LIPOGLY-02 · v2
The question is whether giving a single long-acting antibiotic shortens how long patients spend in hospital compared with usual care. Only a handful of studies here actually compare hospital time against another group, and none of them randomly assigned patients, so the groups being compared may differ in ways that have nothing to do with the drug. One cost study [8] found dalbavancin patients would have needed many more hospital days under standard care, but that figure is an estimate of what would have happened, not a measured difference in real discharge dates. The one study that set out to measure length of stay directly [14] found a shorter stay with oritavancin that did not reach statistical significance (2.12 vs 2.59 days, p=0.097). Other cohorts [7] [9] describe short stays but have no standard-care group to compare against. The evidence all leans the same direction but rests on uncontrolled, potentially biased data, which is why the panel treated it as supportive but far from settled.
Two retrospective cohorts bear directly on hospital stay against a comparator. In [8], a cost study of 87 persons who inject drugs, the dalbavancin group would have had a median of 14 additional days of hospitalisation if they had stayed to complete therapy (P=0.014), alongside lower total cost ($31,698.00 vs $45,093.50; P=0.035). In [14], 107 patients, the primary endpoint of average length of stay was a non-significant 2.12 versus 2.59 days (p=0.097) for oritavancin versus a comparator. [7] describes a 5.5-day median length of stay among the 12.8% of 125 dalbavancin patients later admitted but has no standard-care arm, and [9] reports ED length of stay of 5 versus 25 hours (P<0.05) comparing two dalbavancin delivery pathways rather than versus standard care. No randomised trial in the corpus measures hospital length of stay. The contested point: The evidence divides on whether a real reduction in hospital stay has been demonstrated: the only statistically significant comparative figure [8] is a modelled counterfactual from a cost study, not an observed discharge-date difference, while the single study designed around length of stay [14] was explicitly non-significant (p=0.097). Whether these two uncontrolled cohorts, each for a different drug, support a class-wide claim versus standard care is the core point of contention.
All 20 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Novel Antibiotics, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
No randomised controlled trial in the corpus measures hospital length of stay versus standard care; the entire comparative basis is two retrospective cohorts ([8], [14]) exposed to selection bias and secular trends from different enrollment periods.
The only significant result [8] is a counterfactual ('would have had... if they had stayed to complete their therapy') derived from protocol-mandated standard-of-care durations in a cost study, not an observed difference in actual discharge dates.
Where it remains uncertain. The only significant comparative result [8] is a derived counterfactual rather than an observed difference, and the one dedicated length-of-stay endpoint [14] did not reach significance.
Aleatory
Comparative estimates come from small samples (87 in [8], 107 in [14]), so random variation alone leaves the true stay difference imprecise.
Epistemic
No trial in the corpus randomises patients to a long-acting lipoglycopeptide versus standard care with hospital stay as an outcome, so the causal effect on stay is unknown.
Inferential
[8]'s significant figure is a modelled estimate of days that would have occurred under standard care, not an observed difference, so inferring a real stay reduction from it is a leap.
The load-bearing papers from a corpus of 20, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Only significant comparative signal of reduced hospital days for a long-acting lipoglycopeptide versus standard of care, though it is a derived counterfactual within a retrospective cost study.
“Compared to the SOC group, the DBV group would have had a median of 14 additional days of hospitalization if they had stayed to complete their therapy (P = 0.014).”
Panel auditInternal validity: medium · scope of inference: Hospitalised persons who inject drugs with severe gram-positive infections, on cost and hospital days.Key limitations: retrospective historical control · counterfactual estimate not observed stay · single injecting-drug-use population
Shorter ED length of stay under a dalbavancin telehealth pathway, but the comparison is between two dalbavancin delivery models rather than versus standard care.
“Median ED LOS was significantly shorter in the TC (5 vs 25 hours, P < 0.05).”
Panel auditInternal validity: low · scope of inference: Adults given dalbavancin for ABSSSI across three urban EDs, on recurrence and ED stay by selection criteria.Key limitations: wrong contrast for claim · small retrospective · ED stay only
The one study with length of stay as its primary endpoint; direction favours oritavancin but the result did not reach significance.
“The primary clinical endpoint showed a non-significant decrease in aLOS between the oritavancin group versus comparator (2.12 days versus 2.59 days; p = 0.097).”
Panel auditInternal validity: medium · scope of inference: Patients readmitted for cellulitis or abscess after outpatient antibiotic failure in one community hospital.Key limitations: non-significant result · retrospective, unbalanced groups · post-OPAT-failure population
Descriptive evidence of low admission and short stay after dalbavancin, but with no standard-care comparator it cannot support a reduction claim on its own.
“A total of 16 patients (12.8%) were subsequently admitted to the hospital with a median length of stay of 5.5 days (3.0-8.0).”
Panel auditInternal validity: low · scope of inference: Adults given dalbavancin for ABSSSI in one rural emergency department, on readmission and admission rates.Key limitations: single-arm, no comparator · retrospective · single rural ED
4 of 20 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
A favourable but uncontrolled, confounded signal with one significant counterfactual metric and one non-significant primary length-of-stay endpoint supports an Uncertain verdict rather than a stronger tier.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
No randomised length-of-stay evidence. Every comparative data point here is retrospective and confounded by indication; a randomised design is needed to isolate the drugs' effect on stay. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: A randomised trial allocating eligible patients to a long-acting lipoglycopeptide or standard care, with hospital length of stay as a prespecified primary endpoint.
Open question: Do long-acting lipoglycopeptides shorten actual hospital length of stay versus standard care when patients are randomised? Why it is still open: No randomised trial in the corpus measures hospital stay; both comparative cohorts [8] [14] are retrospective and confounded.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19