Table 3 Mortality HRs from Cox regression by MI.

From: The association between subjective–objective discrepancies in sleep duration and mortality in older men

Predictor

Overall deaths, no. (%)

HR (95% CI)

Unadjusted

Age adjusted

Model 1a

Model 2b

MI as continuous variable

1596 (59.7)

0.57 (0.48–0.67)

0.65 (0.55–0.77)

0.67 (0.57–0.79)

0.69 (0.56–0.84)

MI as categorical variable

Q4 (> 0.08)

374 (57.9)

1.05 (0.93–1.19)

1.04 (0.92–1.17)

1.01 (0.90–1.15)

0.97 (0.85–1.11)

IQR (−0.19 to 0.08)

750 (55.6)

Reference

Reference

Reference

Reference

Q1 (≤ −0.19)

472 (69.6)

1.51 (1.35–1.70)

1.37 (1.22–1.54)

1.38 (1.23–1.55)

1.28 (1.12–1.46)

  1. MI misperception index.
  2. aModel 1 included age, race (white or non-white), education, body mass index, smoking status, weekly alcohol consumption, daily caffeine use, antidepressants, benzodiazepines, sleep medications, and site.
  3. bModel 2 included Model 1 plus overall arousal index, percentage of sleep time oxygen saturation below 80%, Stage 2%, Stage 3–4%, Stage REM%, Epworth Sleepiness Scale score, Pittsburgh Sleep Quality Index, Teng Mini-Mental State Examination score, Physical Activity Scale for the Elderly score, Geriatric Depression score, congestive heart failure, chronic obstructive pulmonary disease, diabetes, heart attack, stroke, wake after sleep onset, objective total sleep time, hypertension, apnea hypopnea index, periodic limb movements index, habitual bed time, habitual wake-up time, and subjective sleep wellness for the PSG night.
Back to article page