Salt the Food, Nudge the Memory — but Only in Men (for Now)

Higher sodium intake was associated with a small but statistically significant increase in the rate of episodic memory decline in older men — but not women — over 6 years. Notably, this relationship persisted after adjusting for systolic and diastolic blood pressure, suggesting blood pressure alone may not fully account for the association.

The study followed 1,208 cognitively unimpaired Australians (mean age ~71 years, 41% male) from the AIBL cohort across 4 assessment time points over 72 months. Baseline sodium intake — captured via food frequency questionnaire — averaged 1,814mg/day in men versus 2,036 mg/day in women.

Linear mixed models showed a statistically significant association between higher sodium intake and faster decline in episodic recall in men (β = −0.00002; FDR-adjusted p = 0.044). No significant associations were observed in women, and no other cognitive domains were affected. In the cohort overall, sodium intake was not significantly associated with cognitive decline (a borderline trend was observed but did not reach statistical significance).

Here’s the interesting wrinkle: sensitivity analyses adjusting for systolic and diastolic blood pressure — individually and together — did not materially change the finding. That doesn’t rule out vascular mechanisms, but it suggests blood pressure alone may not fully explain the association.

“Further investigation, including sex-specific approaches, is required to evaluate how sodium intake could be incorporated as one modifiable lifestyle factor aimed at delaying Alzheimer’s disease onset.” — Chuwa et al.

The proposed mechanisms come from prior literature, not this study directly: excess sodium has been linked to neuroinflammation, impaired endothelial nitric oxide signalling, blood–brain barrier disruption, and reduced cerebral perfusion — processes that could plausibly affect hippocampal-dependent memory. Animal studies also implicate increased hyperphosphorylation and tau aggregation, though this remains untested in humans.

Clinical takeaway: This is an observational study using a single baseline dietary assessment (which did not capture discretionary salt), in a predominantly white Australian cohort — so it doesn’t establish causality and shouldn’t be overgeneralized.

But for older male patients already being counseled on cardiovascular risk, sodium reduction now adds a tentative, hypothesis-generating cognitive consideration that may be worth mentioning — with appropriate caution.

Source: Neurobiology of Aging

A Second Pregnancy Leaves a Different Mark on the Brain Than the First One Did

A machine learning classifier, fed only MRI-derived brain change maps, correctly identified which women had undergone a first versus second pregnancy 80% of the time (leave-one-out cross-validation; 70% with the more conservative k-fold approach, both significant) — suggesting these reflect distinguishable patterns of neural change, rather than identical processes repeating.

Both pregnancies produced widespread gray matter volume reductions — roughly 2.8% in second-time mothers versus 3.1% in first-time mothers — but the location of change diverged. First pregnancies showed more extensive and pronounced changes in the default mode and frontoparietal networks, the circuitry tied to introspection and higher-order cognition. Second pregnancies showed relatively greater changes in the dorsal attention and somatomotor networks, along with a decrease in mean diffusivity of the right corticospinal tract that remained evident up to a year postpartum in a subset of participants.

The sneaky part: Default mode network coherence — which increased across a first pregnancy — did not show the same increase in second-time mothers. That functional shift appears to be a first-pregnancy-dominant adaptation rather than a repeated effect.

"Both a first and second pregnancy particularly strongly impacted the introspective default mode network and the frontoparietal network… However, these changes were more prominent in a first pregnancy, suggesting a primary adaptation of this network in women who become mothers for the first time that is further fine-tuned during a second pregnancy." — Straathof et al., Nature Communications, 2026

The mechanisms remain unclear — MRI cannot resolve the underlying cellular processes. The authors speculate that the externally oriented network changes in second-time mothers may relate to the demands of caring for multiple children, though this is explicitly framed as a hypothesis rather than a demonstrated mechanism.

Brain volume changes were associated with measures of peripartum depression and psychological distress in both groups; these associations were more widespread during pregnancy in second-time mothers and during the postpartum period in first-time mothers, though causality cannot be inferred.

Clinical bottom line: The maternal brain is not simply “already changed” after a first pregnancy — a second one is associated with both overlapping and distinct patterns of structural and functional change. The earlier emergence of depression-related brain associations in multiparous women (during pregnancy rather than postpartum) is an interesting signal, but remains exploratory and requires replication.

Source: Nature Communications

The Drug Everyone Calls “Immunosuppressive” Might Be Misunderstood in the ICU

Corticosteroids are routinely labeled as immunosuppressive. But in critical care, that assumption may be too blunt to be useful.

A new review from Sorbonne University, Rayan Braïk revisits a longstanding clinical tension: if steroids suppress immunity, why don’t they seem to increase infections in ICU patients—and in some cases, may even be associated with fewer?

Across trials in septic shock, ARDS, and community-acquired pneumonia, corticosteroids have not been linked to higher rates of secondary infections. A 2019 Cochrane meta-analysis covering more than 5,000 patients found no significant increase in superinfections. Other studies have reported something more counterintuitive: trauma patients treated with hydrocortisone developed less secondary pneumonia, and surgical patients receiving dexamethasone had lower rates of sepsis at discharge.

The review doesn’t claim to resolve this paradox. Instead, it asks whether the underlying premise—that corticosteroids are straightforwardly immunosuppressive in acute illness—is too simplistic.

Part of the answer may be dose and duration. The regimens used in intensive care—typically 0.5 to 2 mg/kg for less than 2 weeks—are far lower than the high-dose “pulse” therapies used in autoimmune disease, where infection risk is well established. But the paper argues that pharmacology alone doesn’t explain the discrepancy.

A more important factor may be timing.

In sepsis and other acute inflammatory states, the immune response appears to follow a biphasic course: an early pro-inflammatory surge, followed by a more prolonged immunosuppressive phase. One hypothesis is that the depth of that later immunosuppression reflects the intensification of the initial inflammatory spike. If so, dampening the early phase with low-dose corticosteroids could theoretically reduce—not worsen—the downstream immune deficit.

This remains a conceptual model, not a settled mechanism. The underlying biology is complex and still evolving. Corticosteroids are known to suppress pro-inflammatory signaling, but they may also preserve functions like neutrophil phagocytosis, alter macrophage metabolism (including itaconate production), and potentially allow for immune “rebound” after discontinuation. How these effects integrate over time is not yet clear.

What is clearer is that the binary label may not fit.

“In the context of acute inflammatory conditions in critical care, classifying corticosteroids as immunosuppressive agents is not biologically justified,” Braïk writes. “It would be more accurate to refer to them as immunomodulators.”

Clinically, that reframing has implications. For conditions like ARDS, septic shock, and severe pneumonia, concern about infection risk with short-course, low-dose steroids may be less supported than often assumed. The more relevant variable, the paper suggests, may be alignment with the patient’s immunological phase—treating early enough to blunt harmful inflammation, but not so long as to extend into the period of immune suppression.

The challenge is that this phase isn’t routinely measured.

The review argues for greater use of immunomonitoring—tracking markers like IL-6, IL-10, or HLA-DR expression—to better time therapy. Until then, corticosteroids in the ICU may be less about suppressing immunity than about reshaping it—depending on when, how, and how long they’re used.

Source: Anaesthesia Critical Care & Pain Medicine

The clinical literature. Applied to the patients in your waiting room.