Vasoactive-Inotropic Score (VIS)

VIS = dopamine + dobutamine + 100×epinephrine + 10×milrinone + 10,000×vasopressin + 100×norepinephrine — the original six-drug weighted sum of Gaies et al. 2010. Every rate is in mcg/kg/min except vasopressin, which is in units/kg/min (coefficient 10,000). Each drug is optional; an agent not running contributes 0. There are no branches, floor, ceiling, or age adjustment: the result is a single continuous index (displayed to one decimal place) with no interpretation bands applied.

Original six-drug Gaies 2010 VIS = dopamine + dobutamine + 100×epinephrine + 10×milrinone + 10,000×vasopressin(units/kg/min) + 100×norepinephrine. All six coefficients are confirmed verbatim against Box 1 of the Gaies 2010 primary and are unchanged in the Gaies 2014 re-derivation. Phenylephrine is not part of VIS — it appears in neither Gaies 2010 nor Gaies 2014 — so its absence here is correct and deliberate, not a gap. Nor are any newer agents offered, and that is now a positive decision rather than an omission: the coefficients proposed for them disagree sharply between proposals — methylene blue 1 versus 20, angiotensin II 0.25 versus 25, olprinone 10 versus 25 — so any calculator including them would have to pick a winner between two figures differing by up to a hundredfold and would stop being comparable to the published literature. The levosimendan (×50) and phenylephrine (×10) terms of some adult and ECMO variants are excluded for the same reason of fidelity, so the output is always a true Gaies VIS. The provenance of these left-out coefficients is not uniform, so it is stated per term rather than as one blanket caveat. The three newer-agent pairs above are quoted from the 2026-08-04 review, and the competing proposals behind them are [NEEDS SOURCE] for a primary. Levosimendan ×50 is not unsourced at all: it was confirmed in the 2026-07-25 verification pass against the independently fetched full texts of two variant cohorts, an ECMO one and a heart-transplant one. Phenylephrine ×10 comes from that same 2026-07-25 pass, corroborated by two independent secondary aggregations of modified-VIS formulas but still [NEEDS SOURCE] for a directly fetched primary. Nothing here computes any of these coefficients, so every one of these gaps is descriptive rather than load-bearing. VIS is a continuous index of vasoactive/inotropic support intensity, not a diagnostic test or clinical device, and higher values are an association marker for morbidity/mortality in the cited cohorts, not a treatment trigger. A threshold and an effect size only mean something together, and both published Gaies pairings can now be stated in full. Gaies 2010 (adjusted OR 8.1, 95% CI 3.4–19.2, p<0.001, high versus low VIS): Table 1 on page 235 was read directly on 2026-08-04, and it sorts patients into five groups by the score reached in each of two periods — group 1 is under 10 in the first 24h and under 5 in hours 24–48; group 2 is 10–14 and 5–9; group 3 is 15–19 and 10–14; group 4 is 20–24 and 15–19; group 5 is 25 or more and 20 or more. Each patient is assigned to the highest group reached in EITHER period (the table's own illustration puts a patient with a maximum of 22 in the first 24h and 14 in the next into group 4), and groups 4 and 5 combined form the high-VIS arm. High VIS therefore means a maximum VIS of 20 or more in the first 24h OR 15 or more in hours 24–48. Two readings must not be carried forward. A literature review reports this as 'more than 15 in the first 24h'; that is wrong — 15–19 in the first period is group 3, which sits in the low-VIS arm, so that reading would move a whole group across the dichotomy. And this calculator's own earlier note, having correctly reconstructed the dual-period rule from a peer-reviewed paraphrase, then demoted it on 2026-08-03 as merely the right neighbourhood with the wrong shape when the five-group scheme was read but its table was not; the printed table shows the dual-period reconstruction was right all along, and it is restored rather than silently re-reversed. The paper's anchoring of the scheme at a VIS of about 15 as the first-period midpoint is consistent with this: 15–19 is the middle group of the five. Gaies 2014 (adjusted OR 6.5, 95% CI 2.9–14.6): a single flat cut-point, maximum VIS of 20 or more during the first 24h, on the same six coefficients. It remains the simpler bedside form, but the 2010 rule is no longer the unstateable one. No cut-point is applied automatically by this calculator, and the reason is that the wider literature has not converged on one: optimal cut-points reported across studies span roughly 10 to 30 with no agreed value (Belletti 2021), Davidson 2012 gives a cohort-specific VIS-at-48h threshold of 10.5 in neonates/infants after cardiac surgery, an ECMO cohort gives a pre-ECMO 61.4, and paediatric septic shock has been reported with a cut of 11 (sensitivity 78.87%, specificity 72.22%, AUC 0.779 — figures from the 2026-08-04 review, whose primary derivation study is not named and is [NEEDS SOURCE] here). That spread is the point: cut-points do not transfer between populations — post-cardiac-surgery, sepsis and ECMO cohorts each derive their own — and the literature says so explicitly. VIS is a snapshot; the prognostic quantity in the literature is the maximum over a defined window (two 24h periods in Gaies 2010, the first 24h in Gaies 2014, 48h in Davidson 2012), which the platform must define and label. Per-drug dose ceilings: none are published for VIS — searched and confirmed absent, not merely unfound — so the maxima this calculator enforces are a local input-validity convention and carry no clinical authority. Vasopressin unit trap: it is dosed in units/kg/min (coefficient 10,000), the only agent not in mcg/kg/min.