Effect of Regional vs General Anesthesia on Incidence of Postoperative Delirium in Older Patients Undergoing Hip Fracture Surgery

**Intervention:** Regional anesthesia (spinal block, epidural, or both techniques combined) with no sedation. Patients were awake or lightly drowsy but not given sedative drugs during surgery.

**Comparator:** General anesthesia (intravenous drugs, inhaled gases, or a combination) where patients were fully unconscious with a breathing tube or mask.

**Primary outcome:** Incidence of delirium during the first 7 days after surgery, assessed twice daily using the Confusion Assessment Method (CAM).

**Secondary outcomes:** Delirium severity (scored on the Memorial Delirium Assessment Scale, range 0–30, higher = worse), delirium duration (days), delirium subtype (hyperactive, hypoactive, mixed), postoperative pain (numeric rating scale 0–100), length of hospital stay, 30-day all-cause mortality, and adverse events (nausea, vomiting, hypotension, etc.).

**Sample size:** 950 patients randomized (476 regional anesthesia, 474 general anesthesia). 941 were evaluable for the primary outcome (6 canceled surgery, 3 withdrew consent).

**Population:** Adults aged 65 years and older (mean age 76.5 years) with a fragility hip fracture requiring surgical repair. 26.8% were male. Patients with or without preexisting dementia were included.

**Setting:** 9 university teaching hospitals in Southeastern China. Enrollment from October 2014 to September 2018; 30-day follow-up ended November 2018.

**Exclusion criteria:** Patients with contraindications to regional anesthesia (e.g., bleeding disorders, infection at injection site), severe cognitive impairment that prevented informed consent (though patients with dementia could be enrolled via proxy consent), or multiple fractures.

**Delirium diagnosis:** Confusion Assessment Method (CAM) — a validated, structured interview that assesses four features: acute onset and fluctuating course, inattention, disorganized thinking, and altered level of consciousness. Administered twice daily (morning and evening) for the first 7 postoperative days by trained research nurses.

**Delirium severity:** Memorial Delirium Assessment Scale (MDAS) — a 10-item clinician-rated scale (0–30, higher = more severe). Administered once daily if delirium was present.

**Delirium subtype:** Classified as hyperactive (agitation, restlessness), hypoactive (lethargy, reduced movement), or mixed (both) based on the Richmond Agitation-Sedation Scale (RASS) and clinical observation.

**Pain:** Numeric Rating Scale (0–100, 0 = no pain, 100 = worst imaginable pain). Assessed at rest and during movement at 6, 12, 24, 48, and 72 hours post-surgery.

**Mortality:** All-cause death within 30 days of surgery, confirmed by hospital records or phone follow-up.

**Adverse events:** Systematically recorded using a predefined checklist of common postoperative complications (nausea, vomiting, hypotension, respiratory depression, cardiac events, etc.).

**Study design:** This was a randomized, allocation-concealed, open-label, multicenter clinical trial. Patients were randomly assigned (1:1 ratio) to receive either regional anesthesia or general anesthesia using a computer-generated randomization sequence with variable block sizes (4, 6, or 8). Allocation was concealed using sequentially numbered, opaque, sealed envelopes — meaning the person enrolling the patient could not predict which group the patient would enter.

**Blinding:** The trial was open-label — patients, surgeons, anesthesiologists, and research nurses knew which anesthesia type was given. This is a major limitation because delirium assessment (CAM) requires subjective judgment, and knowledge of the treatment could bias the assessor. The authors attempted to mitigate this by having the outcome assessors (research nurses) not be involved in the patient's clinical care, but they were still aware of the group assignment. There was no sham procedure (e.g., a placebo spinal injection for the general anesthesia group), so blinding was impossible for practical and ethical reasons.

**Duration:** The intervention was a single surgical event (hip fracture repair). Follow-up for the primary outcome was 7 days post-surgery. Secondary outcomes were assessed up to 30 days (mortality) or hospital discharge (length of stay). This is a short follow-up window — delirium can occur beyond 7 days, and longer-term cognitive outcomes were not measured.

**Statistical approach:** The primary analysis was intention-to-treat (all randomized patients analyzed in their assigned group, regardless of whether they received the planned anesthesia). The primary outcome (delirium incidence) was compared using a χ² test, with risk difference and relative risk reported with 95% confidence intervals. A pre-specified subgroup analysis examined delirium risk by age (<75 vs ≥75), sex, dementia status, and fracture type. The sample size was calculated to detect a 7% absolute reduction in delirium (from 15% to 8%) with 80% power at α=0.05 — but the actual delirium rate was much lower than expected (5–6%), meaning the study was underpowered to detect smaller but still clinically meaningful differences.

**What this design can and cannot prove:**

**Can prove:** Because of randomization, the study can establish that the choice of anesthesia type (regional vs general) does not cause a large difference in delirium risk in this specific population. The groups were well-balanced at baseline (age, sex, dementia prevalence, fracture type), so confounding is unlikely to explain the null result.

**Cannot prove:** The open-label design means we cannot rule out bias in delirium assessment — if nurses expected regional anesthesia to reduce delirium, they might have unconsciously rated those patients differently. The study cannot tell us about long-term cognitive effects (beyond 7 days) or about delirium risk in other surgical populations (e.g., cardiac surgery, major abdominal surgery). It also cannot tell us whether regional anesthesia with sedation (which is common in practice) would have different effects — the protocol specifically prohibited sedation in the regional group, so these results apply only to "awake" regional anesthesia.

**Major methodological weaknesses:**

1. **Open-label design** — no blinding of patients, clinicians, or outcome assessors.

2. **Lower-than-expected delirium rate** — the study was powered for 15% delirium but observed ~5.5%, making it underpowered to detect modest but clinically relevant differences.

3. **Short follow-up** — 7 days for delirium may miss late-onset cases.

4. **Single country** — all sites in China; generalizability to other healthcare systems and populations is uncertain.

5. **No sham control** — the regional group received a spinal/epidural injection, while the general group received intravenous drugs; the psychological experience of being awake vs asleep during surgery could influence delirium independently of the anesthetic drugs.

**Primary outcome — delirium incidence (first 7 days):**

- Regional anesthesia: 29/470 (6.2%)

- General anesthesia: 24/471 (5.1%)

- Unadjusted risk difference: 1.1% (95% CI, -1.7% to 3.8%; p=0.48)

- Unadjusted relative risk: 1.2 (95% CI, 0.7 to 2.0; p=0.57)

- **Conclusion:** No statistically significant difference. The confidence interval includes the possibility of regional anesthesia increasing delirium by up to 3.8% or decreasing it by up to 1.7%.

**Delirium severity (MDAS score, mean):**

- Regional: 23.0 vs General: 24.1 (difference -1.1; 95% CI, -4.6 to 3.1) — not significant.

**Delirium duration (days, mean):**

- Regional: 2.3 vs General: 2.0 (difference 0.3; 95% CI, -0.7 to 1.3) — not significant.

**Delirium subtype (among those with delirium):**

- Hypoactive: Regional 11/29 (37.9%) vs General 5/24 (20.8%) — risk difference 11.5% (95% CI, -11.0% to 35.7%) — not significant, but a trend toward more hypoactive delirium with regional anesthesia.

- Hyperactive: Regional 5/29 (17.2%) vs General 7/24 (29.2%)

- Mixed: Regional 13/29 (44.8%) vs General 12/24 (50.0%)

**Postoperative pain (worst pain score, median):**

- Regional: 0 (IQR 0–20) vs General: 0 (IQR 0–10) — difference 0 (95% CI, 0 to 0). Pain was very low in both groups.

**Length of hospitalization (median):**

- Regional: 7 days (IQR 5–10) vs General: 7 days (IQR 6–10) — no difference.

**30-day all-cause mortality:**

- Regional: 8/470 (1.7%) vs General: 4/471 (0.9%) — risk difference -0.8% (95% CI, -2.2% to 0.7%; RR 2.0, 95% CI 0.6 to 6.5) — not significant, but the wide confidence interval means a doubling of mortality risk cannot be ruled out.

**Adverse events (total episodes):**

- Regional: 106 vs General: 102

- Nausea/vomiting: Regional 47 (44.3%) vs General 34 (33.3%) — more common with regional.

- Postoperative hypotension: Regional 13 (12.3%) vs General 10 (9.8%)

- Other events (cardiac, respiratory, renal) were rare and similar between groups.

**Subgroup analyses (pre-specified):** No significant interaction between anesthesia type and age, sex, dementia status, or fracture type. The null result was consistent across all subgroups.

The difference in delirium risk between regional and general anesthesia was tiny — about 1 percentage point in absolute terms. To put this in perspective: if you treated 100 patients with regional anesthesia instead of general anesthesia, you would expect roughly 1 more case of delirium (not fewer). The 95% confidence interval tells us the true effect could be anywhere from a 1.7% reduction to a 3.8% increase in delirium — meaning regional anesthesia could be slightly beneficial, slightly harmful, or neutral. Given the low baseline delirium rate (5.1%), even a best-case scenario would prevent only about 2 cases per 100 patients.

The severity scores were nearly identical (23 vs 24 on a 30-point scale), and the duration of delirium was the same (~2 days). Pain control was excellent in both groups (median pain score of 0). Length of stay was identical (7 days). Mortality was numerically higher with regional anesthesia (1.7% vs 0.9%), but this was not statistically significant and could be due to chance.

In plain English: **Choosing regional over general anesthesia does not meaningfully reduce your risk of postoperative delirium.** If anything, the data hint at a possible small increase in risk, but the study was too small to be sure. The practical difference between the two approaches for delirium prevention is essentially zero.

**What the authors acknowledge:**

Open-label design could introduce bias in delirium assessment.

Lower-than-expected delirium rate reduced statistical power.

The protocol prohibited sedation in the regional group, which may not reflect real-world practice where some sedation is often used.

Results may not generalize to non-Chinese populations or to other types of surgery.

**What a critical reader would note:**

**No blinding of outcome assessors:** The research nurses who administered the CAM knew which anesthesia the patient received. Since CAM requires subjective interpretation (e.g., "disorganized thinking"), this is a serious source of potential bias. A blinded assessment (e.g., using video recordings rated by independent clinicians) would have been stronger.

**Short follow-up:** Delirium can occur up to 30 days post-surgery, especially in frail older adults. The 7-day window may miss late-onset cases. The study also did not measure long-term cognitive decline, which is a more clinically meaningful outcome than transient delirium.

**Single country, specific healthcare system:** All hospitals were in Southeastern China. Differences in perioperative care (e.g., nursing ratios, pain management protocols, family involvement) could affect delirium rates and limit generalizability to Western settings.

**No sham control:** Patients in the regional group experienced being awake during surgery, which could be psychologically distressing and itself increase delirium risk. The general group had no such experience. This confounds the anesthesia type with the psychological experience of surgery.

**Multiple secondary outcomes without correction:** The authors reported many secondary outcomes (delirium severity, duration, subtype, pain, mortality, complications) without adjusting for multiple comparisons. Some of the observed differences (e.g., more hypoactive delirium with regional) could be due to chance.

**Industry funding not reported:** The trial was funded by the Chinese government (National Natural Science Foundation), so industry bias is unlikely, but the lack of commercial interest also means there was no incentive to design a more rigorous blinded study.

**Exclusion of severe dementia:** While patients with dementia were included, those with severe cognitive impairment who could not provide consent (and had no proxy) were excluded. This limits generalizability to the most vulnerable patients, who are at highest risk for delirium.

For someone running their own n=1 experiment (e.g., if you are an older adult facing hip surgery and want to reduce your delirium risk):

### What to test

**Do not test anesthesia type.** This study shows that choosing regional vs general anesthesia is unlikely to meaningfully change your delirium risk. Instead, focus on modifiable factors that have stronger evidence: minimizing sedative drugs (especially benzodiazepines), optimizing pain control with non-opioid alternatives, maintaining sleep-wake cycles, early mobilization, and avoiding dehydration and electrolyte imbalances.

If you are determined to test anesthesia type, the intervention would be "request regional anesthesia without sedation" vs "accept general anesthesia." But the data suggest this is a low-yield experiment.

### Minimum meaningful duration

Delirium typically occurs within the first 3–5 days after surgery, but can appear up to 30 days. For a self-experiment, monitor for at least 7 days post-surgery, ideally 14 days. Use twice-daily assessments (morning and evening) to capture fluctuating symptoms.

### What to measure

**Primary metric:** Presence of delirium — use a validated screening tool like the 3-Minute Diagnostic Interview for CAM (3D-CAM) or the Brief Confusion Assessment Method (bCAM). These are free, take <5 minutes, and can be administered by a family member or caregiver after brief training.

**Secondary metrics:**

- Delirium severity (e.g., MDAS or Delirium Rating Scale-Revised-98)

- Sleep quality (e.g., sleep diary or wrist actigraphy)

- Pain scores (0–10 scale, at rest and with movement)

- Medication use (especially benzodiazepines, anticholinergics, opioids)

- Hydration status (urine output, skin turgor)

- Mobility (time to first walk, steps per day)

**Confounders to track:** Age, baseline cognitive function (use Montreal Cognitive Assessment or Mini-Mental State Exam), pre-existing sleep disorders, alcohol