NEJM This Week — August 6, 2026

A strong week for the generalist. LOGICAL is the trial to know — conservative oxygen after cardiac arrest did not improve survival with a favorable functional outcome, closing out a decade of post hoc subgroup hope and joining the other neutral post-arrest target trials. The Syncope Clinical Practice review is unusually practical, and its best material is about older adults who fall: carotid sinus syndrome explains up to 30% of otherwise-unexplained falls, and urinary incontinence does not separate syncope from seizure. A superb Clinical Problem-Solving case lands on checkpoint inhibitor–associated HLH — with a timing rule that cleanly separates it from cytokine release syndrome. Two trials are summarized for awareness: PROTEUS (perioperative apalutamide around prostatectomy) and SirPAD (sirolimus-coated balloons for peripheral artery disease).

Critical CareLOGICAL · RCTHigh yield · IM

Conservative Oxygen for Unresponsive Patients after Cardiac Arrest

In unresponsive, mechanically ventilated ICU patients after cardiac arrest, does conservative vs liberal oxygen therapy improve survival with a favorable functional outcome at 180 days?

Population: 1840 adults enrolled (ITT 1821: 873 vs 948) from 53 ICUs in Australia, New Zealand, and Ireland. Mechanically ventilated after in- or out-of-hospital arrest, unable to follow commands with clinical concern for ischemic encephalopathy; few exclusions. Open-label, assessor-blinded; nested within Mega-ROX.
Intervention: Conservative oxygen — upper-limit Spo₂ alarm at 95%, Fio₂ decreased to 0.21 whenever Spo₂ was above the lower limit.
Comparison: Liberal oxygen — no upper Spo₂ limit; minimum Fio₂ 0.3 during ventilation. The lower Spo₂ alarm was 90% in both arms.
Outcome: Favorable functional outcome (GOS-E 5–8) at 180 d: 38.2% vs 39.7% (adjusted absolute risk difference −0.9 pp, 95% CI −5.5 to 3.7; RR 0.97, 95% CI 0.87–1.09; P=0.65). Alive at 180 d 48.0% vs 49.7%. Quality of life and cognition similar. No adverse events reported.

Clinical takeaway. A pragmatic, generalizable, decisively neutral trial — and the editorial (Andersen & Granfeldt) supplies the caveat that keeps it honest: the intervention was delayed (median 7 h from ROSC to randomization vs ~2 h in comparable trials) and the separation between arms was small (median Fio₂ 0.25 vs 0.30; median Pao₂ 74 vs 86 mm Hg), with few patients reaching severe hyperoxemia. So this tests modest oxygen titration in the ICU, not avoidance of the extreme early hyperoxia that harms animals. Protocol deviations were real: 26.3% of the conservative arm failed to drop Fio₂ when Spo₂ was ≥95%. What this changes: confirms BOX and overturns the earlier post hoc subgroup signals — LOGICAL joins the neutral trials of oxygen, CO₂, blood-pressure, and temperature targets, so post-arrest supportive care should not differ from that of other ICU patients until something is proven superior.

Heme / OncologyPROTEUS · RCTNotable · narrow onc

Perioperative Apalutamide in High-Risk Localized Prostate Cancer

In high-risk localized or locally advanced prostate cancer treated with radical prostatectomy, does perioperative ADT plus apalutamide vs ADT plus placebo improve pathological response and metastasis-free survival?

Population: 2109 surgically eligible men (1057 vs 1052) with newly diagnosed high-risk localized or locally advanced (pelvic node–positive) prostate cancer. Median follow-up 61.7 months. Phase 3, double-blind, placebo-controlled.
Intervention: Apalutamide 240 mg daily + ADT for 6 cycles (28 d each) before and after radical prostatectomy with pelvic lymph-node dissection (~12 months total).
Comparison: ADT + placebo on the same schedule. Adjuvant or salvage radiotherapy permitted in both arms.
Outcome: Both primary endpoints met. Pathological complete response or minimal residual disease 8.9% vs 1.0% (OR 10.17; 95% CI 5.27–19.64; P<0.001). 5-yr metastasis-free survival 78.2% vs 73.5% (HR 0.80; 95% CI 0.67–0.96; P=0.02). Grade 3/4 adverse events 39.6% vs 31.0%, driven by rash (21.2% vs 10.0%); hot flush 63.4% vs 56.5%.

Clinical takeaway. The editorial (Antonarakis, "A Watershed Moment") explains why this matters: guidelines currently endorse radical prostatectomy alone without systemic therapy, precisely because randomized data showing a metastasis-free or overall-survival gain from adding systemic therapy to surgery have been lacking — while roughly 50% recur within 5 years. PROTEUS is the first randomized trial to fill that gap. For the internist the practical residue is the toxicity profile and the fact that these men now arrive on a year of androgen-pathway blockade. What this changes: extends androgen-receptor pathway inhibition from metastatic/castration-resistant disease into the perioperative curative-intent setting — a urologic-oncology decision, not an internist's.

Vascular / CardiologySirPAD · RCTNotable · procedural

Sirolimus-Coated Balloon Angioplasty for Infrainguinal Artery Disease

In infrainguinal peripheral artery disease requiring endovascular treatment, does a sirolimus-coated balloon vs an uncoated balloon reduce major adverse limb events at 1 year?

Population: 1252 patients (626 vs 626); median age 75, 35.1% women. Deliberately inclusive: a substantial share had critical limb ischemia (usually excluded from device trials), ~1/3 infrapopliteal disease, median lesion length 150 mm, >50% occlusions, adjunctive stenting in >30%. Open-label with blinded adjudication; sequential noninferiority→superiority testing.
Intervention: Angioplasty with a sirolimus-coated balloon (cytostatic; the coronary-stent antirestenosis agent).
Comparison: Angioplasty with an uncoated balloon.
Outcome: Primary composite (unplanned major amputation of the target limb or target-lesion revascularization for critical limb ischemia at 1 yr) 8.8% vs 15.0% (risk difference −4.9 pp; 95% CI −8.5 to −1.3; P<0.001 noninferiority, P=0.009 superiority). Key secondary 23.0% vs 30.8% (P=0.002). All-cause death 11.8% vs 12.8% (P=0.67).

Clinical takeaway. The editorial (Armstrong) frames SirPAD as a strategy trial rather than a device-registration trial: enrolling complex, critical-limb-ischemia patients with long occlusive lesions is exactly what earlier paclitaxel trials avoided, and it is what makes this result generalizable. Two things worth carrying: prior trials were powered for restenosis, not for amputation or clinically-driven revascularization, and the historical paclitaxel mortality signal does not reappear here (death 11.8% vs 12.8%). What this changes: extends drug-coated-balloon benefit from a restenosis surrogate to hard limb outcomes and shifts the field from paclitaxel toward sirolimus — a vascular-specialist decision, useful context when your PAD patient is referred.

CardiologyReview · High yield · IM

Syncope

In an older adult with an unexplained fall or transient loss of consciousness, how do you separate reflex, orthostatic, and cardiac syncope — and which findings mandate a cardiac workup?

•The unexplained fall IS the presentation in older adults. Falls account for ~15% of older adults' ED visits, and 20–30% of those falls are unexplained — potentially unrecognized syncope. Carotid sinus syndrome occurs almost exclusively in older adults and explains 9–17% of syncopal events and up to 30% of otherwise-unexplained falls. •Carotid sinus syndrome is defined by a pause of ≥3 seconds OR a systolic BP fall of ≥50 mm Hg during carotid sinus massage. Perform massage both supine and upright with continuous HR/BP monitoring, in patients >40 years with suspected reflex syncope, syncope on head turning, or unexplained falls. •Urinary incontinence does NOT distinguish syncope from epilepsy — it occurs in 10–20% of syncopal events, particularly in older adults. Tongue biting, jerking movements, and post-event drowsiness also occur in syncope. Epilepsy is the most common syncope mimic. •The mix inverts with age: reflex syncope causes >90% of events in younger adults but only ~40% in older adults, while cardiac syncope climbs from ≤5% to ~30%. Complex syncope (several coexisting mechanisms) affects up to a third of older patients. Recurrent syncope carries higher mortality (HR 1.87; 95% CI 1.26–2.77) and MACE (HR 2.69; 95% CI 2.02–3.59) at 24 months. •Red flags for a cardiac cause: age >60, known cardiac/structural/valvular disease, male sex, brief or absent prodrome, palpitations or chest pain beforehand, sudden onset, syncope during exertion or while seated/supine, few prior episodes, abnormal cardiac exam, family history of inheritable disease or sudden cardiac death <50. •Deprescribing is treatment. Among antihypertensives, ACE inhibitors (OR 0.85; 95% CI 0.81–0.89) and calcium-channel blockers (OR 0.81; 95% CI 0.74–0.90) carry the lowest hypotensive-event risk in older adults. Pacing cuts recurrence by >50% in reflex syncope with documented bradycardia; fludrocortisone for vasovagal syncope was only marginally nonsignificant (HR 0.69; 95% CI 0.46–1.03; P=0.07).

Teaching Case · Clinical Problem-Solving

Consumed with Inflammation · Hematology / Oncology

High yield · IM

Key learning point. Final diagnosis: checkpoint inhibitor–associated hemophagocytic lymphohistiocytosis (HLH) after pembrolizumab. A 60-year-old woman with triple-negative breast cancer had a week of fevers, rigors, and cough; the infectious workup was exhaustively negative and she failed vancomycin, cefepime→piperacillin–tazobactam, and doxycycline. The pivot was in the labs: ferritin 37,956 µg/L, triglycerides 296 mg/dL, soluble CD25 4268 pg/mL, IL-6 56.2 pg/mL, splenomegaly on CT, and progressive cytopenias. Methylprednisolone 1 mg/kg produced immediate defervescence.

The discriminator worth memorizing: cytokine release syndrome occurs within days of starting immunotherapy, whereas checkpoint inhibitor–associated HLH occurs later — median 6.7 weeks (range 3–16). Mechanistically, checkpoint blockade disinhibits the costimulatory molecules that normally restrain T-cell activation, releasing IL-6, TNF-α, and IFN-γ. This is not vanishingly rare: a WHO pharmacovigilance database held 681 reports (nivolumab 231, pembrolizumab 179, ipilimumab 169, atezolizumab 90, durvalumab 12). Treat by stopping immunotherapy plus high-dose glucocorticoids, with anakinra or ruxolitinib, and low-dose etoposide if refractory. One diagnostic nuance: HLH-2004 needs 5 of 8 criteria, and a recent study found that dropping NK-cell activity did not reduce diagnostic accuracy.

Also in this issue. Original Article — FIND-CKD (finerenone in CKD without diabetes) appears in this print issue but published online-first on June 4 and was covered in the June 4 digest; its editorial has now published (Toto, "Mineralocorticoid Receptor Antagonism — FINDing Another Benefit"). Perspectives — the inequality–pandemic cycle and preparedness; the social contract of Bundibugyo Ebola isolation; Ebola at 50; "Doctor's Dirty Little Secret — The Cost of Silence." Images in Clinical Medicine — type 1 autoimmune pancreatitis; obstructive hypertrophic cardiomyopathy. Clinical Decisions — Fast Track in Internal Medicine Residency. Correspondence — Bundibugyo virus clinical characteristics (DRC 2026) and cross-reactive antibody responses after licensed Ebola vaccines; cardiotoxic effects after DMD gene therapy; left ventricular unloading in high-risk PCI; nirmatrelvir–ritonavir in higher-risk outpatients; dengue suppression with male Wolbachia-infected mosquitoes.

Summaries are original paraphrases prepared for educational use; figures are drawn from the published full text. Read the full articles at NEJM.org (subscription required). Trial registrations: LOGICAL (ACTRN12621000518864), PROTEUS (NCT03767244), SirPAD (NCT04238546). Citations: N Engl J Med 2026; Vol. 395, No. 6. DOIs: 10.1056/NEJMoa2513814 (editorial 10.1056/NEJMe2609084), 10.1056/NEJMoa2603878 (editorial 10.1056/NEJMe2606250), 10.1056/NEJMoa2600360 (editorial 10.1056/NEJMe2602625), 10.1056/NEJMcp2517255, 10.1056/NEJMcps2516922.