Table 3.

Long-term effects induced by shift work on cognitive functions.

Study Location	Subjects	Length of Employment in Shift Works-Characteristics of the Shift Worker Groups	Cognitive Testing	Results	Quality Rating According to NOS	References
France	3237 workers (1660 M; 1577 F)	Shift-workers: current (265 M; 321 F); former (346 M, 242 F); never (1049 M; 1014 F).	Memory test adapted from the Rey Auditory Verbal Learning Test (RAVLT); Digit Symbol Substitution Test (DSST); Selective attention test derived from the Sternberg test.	Current male shift-workers showed significantly lower cognitive performance in terms of immediate and delayed free recalls, speed measures, compared to never exposed workers. Memory performance: significantly decreased in men according to the 10–20 year shift-work duration). Memory scores were significantly higher for women who stopped 4 years before.	Good	Rouch et al., 2005 [46]
United States	Female nurses: (n = 16,190; mean age = 74.3 ± 2.3 years).	Shift-work experience duration: 1–9 years (n = 7685); 10–19 years (n = 1341); ≥20 years (n = 1028); No shift workers: 6136.	Telephone Interview of Cognitive Status (TICS); Immediate recall and delayed recall of the East Boston Memory Test. Administration: subsequently repeated 3 times, at 2-year intervals.	Mean differences in average cognition were similar for both global and verbal scores in nurses with ≥20 years of shift-work history and no-shift nurses, while a significantly lower cognitive status was assessed through the TICS with respect to no-shift nurses.	Very good	Devore et al., 2003 [48]
France	3119 present and former wage earners covering a wide range of occupations and economic sectors. Shift work current or past exposure: n = 1484; No-shift workers: n = 1635.	Current shift workers (n = 381); ≤5 years recency (n = 417); >5 years recency (n = 295). Rotating shift workers: 1- 10 years (n = 583); ≥10 years (n = 534).	Verbal episodic memory test; Digit-Symbol Substitution test (DSST); Selective attention test derived from the Sternberg test. Administration: the first measurement (t1, 1996), and 5 (t2) and 10 (t3) years later.	Cognitive performance: significantly lower in shift workers vs. controls; global, 56.0 ± 10.71 vs. 53.3 ± 10.60; memory, 50.8 ± 10.61 vs. 48.5 ± 10.46; speed: 78.5 ± 8.77 vs. 76.5 ± 9.05. Rotating shift workers: significantly lower cognitive performance in >10 years workers compared to those engaged for 1–10 years: global, 51.8 ± 10.49 vs. 55.4 ± 10.08; memory, 47 ± 10.23 vs. 50.3 ± 10.33; speed, 75.6 ± 8.84 vs. 78.1 ± 8.2.	Very good	Marquie et al., 2015 [47]
Sweden (2015)	7143 Partecipants (age range = 45–75 years).	Non shift-work (n = 4611); past shift-work (n = 1531); recent former shift-work (n = 358); current shift work (n = 643).	Trail Making Test (TMT A-B). Administration: during the visit to 1 of 2 Swedish test centers.	No significant differences in performance were found in TMT between past and non-shift workers. Current and recent former shift workers required more time to complete the TMT test compared to no-shift workers. The ratio between two subsets of TMT tests, as a more accurate measure of executive functions, was significantly higher in current and recent former shift workers compared to non-shift ones.	Very good	Titova et al., 2016 [31]
Sweden (1986–2002)	Cohort from the Swedish Adoption Twin Study of Aging (SATSA) 1986: n = 595; 299 M, 296 F; mean age = 62.6 ± 8.5 years). Cohort from the Screening Across the Lifespan Twin (SALT) sample (n = 320).	Shift workers (n = 106); years in shift work: 1–9 (40.6%); 10–19 (30.2%); ≥20 (29.2%). SALT-sample: night shift workers.	Verbal ability test; Memory tests. Administration: measures of cognitive performance were assessed in 9 waves during 1986–2012.	Midlife shift work was not associated with mean cognitive performance at retirement age or with a rate of cognitive change during the 27-year follow-up period for any of the cognitive factors (verbal, spatial, memory, processing speed, and general cognitive ability).	Very good	Bokenberger et al., 2017 [27]
Germany(January–April 2017)	500 Healthcare workers (389 F, 111 M; mean age = 61.6 ± 4.2 years). 75 completing a socio- demographic questionnaire; 47 subjects participating in cognitive testing (n = 47; 11% of the total sample).	Mean duration of shift-work experience: 29 and 24 years in any shift system and in a system including night shifts, respectively.	Repeatable Battery for the Assessment of Neuropsychological Status (RBANS); German version of the Trail Making Test (TMT); Letter-Number (LN) span; Vocabulary Test.	Cognitive test scores: in the Vocabulary Test none of the subjects scored below average, in the TMT test (participants’ cognitive performance speed) 17% had a score below average; in the RBANS test (neuropsychological status) 6% were below average. Clinical evaluation: 17 participants showed slight difficulties in at least one subscale of the cognitive tests. The frequency of indication for slight or pronounced impairment did not differ between those working with and without shifts.	Good	Weinmann et al., 2018 [49]
Netherlands	50 Retired maritime pilots (mean age = 71.7 ± 7.7 years).	History of >25 years of work on irregular schedules.	Cognitive Failure Questionnaire (CFQ); Early Dementia Questionnaire (EDQ); Pittsburgh Sleep Quality Index (PSQI); Sleep-wake diaries; QoL (EQ-5D); Hospital Anxiety and Depression Scale (HADS). Administration: participants were scheduled for 1 visit at which they supplied answers to 6 questionnaires.	CFQ: all scores were within the normal range. The highest sub-score was observed on the confusion subscale and the lowest on the orientation subscale. EDQ: all participants remained below the median cut-off median (8), suggesting that participants did not show signs of early dementia.	Good	Thomas et al., 2019 [50]
Netherlands	19 Maritime pilots (median age 53 ± 3.4 years); 16 Controls (mean age = 57 ± 2.9 years).	Work history: an average of 20 years (mean = 19.8; range 10 to 30 years).	Logical Memory Subtest (WMS-IV LM); Rey Auditory Verbal Learning Test (RAVLT)); Letter and semantic fluency, Boston Naming Test; Digit Span subtest (WAIS-IV); Trail Making Test (TMT A-B); WAIS-IV Coding; Test of Attentional Performance (TAP 2.0); Doors Test. Administration: performed in the morning following polysomnography (in 2016 and 2017).	No difference between maritime pilots and controls on episodic memory tests. Small differences in semantic memory and language with slightly better performance for maritime pilots than controls. Working memory, executive function and attention: no significant differences between the groups. All test scores were within normal ranges adjusted for age and education based on available normative data.	Satisfactory	Thomas et al., 2020 [51]