Rotating night shifts too quickly may cause anxiety and decreased attentional performance, and impact prolactin levels during the subsequent day: a case control study

The 59 participants included 23 off-duty (OD) female nurses (mean age 26.1 ± 1.9 years; mean years of education 15.3 ± 1.0), 20 female nurses working two consecutive night shifts (2NS; mean age 26.0 ± 2.0 years, mean years of education 14.7 ± 1.0) and 16 female nurses working four consecutive night shifts (4NS; mean age 27.1 ± 2.0 years; mean years of education 15.4 ± 1.0). All participants worked in the acute ward of Kaohsiung Municipal Kai-Syuan Psychiatric Hospital in southern Taiwan. We excluded those who reported any of the following characteristics on a screening questionnaire: current use of hypnotics, regular coffee-drinking, psychiatric illness, major systemic disease or sleep disorder. Most of the work schedules of our nurses consist of repetitive blocks of two consecutive day shifts (8 a.m. to 4 p.m. or 8 a.m. to 5:30 p.m.), two evening shifts (4 p.m. to 12 a.m.), two night shifts (12 a.m. to 8 a.m.), and then at least 1 day off. Occasionally, nurses work night shifts for 3 to 4 days consecutively because of staffing demands. To prevent adaptation to night shifts, all of the nurses had worked either day shifts or been free of duty for at least 3 days before entering the study. Those in the night shift groups were asked to sleep prophylactically from 7 p.m. to 11 p.m. while working nights. Demographic data including age, years of education, and mean self-reported total sleep time (TST, including daytime sleep and prophylactic sleep) for the night shift groups and the sleep time during the night before the study began for the OD group were recorded. This study was approved by the Ethics Committee of Kaohsiung Municipal Kai-Suan Psychiatric Hospital (KSPH-2011-15) and all participants gave written consent after being fully informed of the nature and procedures of the study.

All participants arrived at the sleep laboratory at about 9:00 a.m. after their two or four consecutive night shifts. On this off-duty day, they spent about 8 h in the laboratory completing a set of measures every 2 h for a total of four times, starting at 9:20 a.m. The measures included the Maintenance of Wakefulness Test (MWT), State-Trait Anxiety Inventory (STAI) [9], Stanford Sleepiness Scale (SSS) [10], Wisconsin Card Sorting Test (WCST) [11], Digit Symbol Substitution Test (DSST), Symbol Search Test (SST) [12], and the modified Multiple Sleep Latency Test (modified MSLT) (Figure 1). Bedside blood samples were collected at the end of the MSLT and were tested for sleep-related hormones. The variation in the time of blood sample collection was within 20 min for each participant and each collection, and occurred at 11:00 a.m.–11:20 a.m., 1:00 p.m.–1:20 p.m., 3:00 p.m.–3:20 p.m., and 5:00 p.m.–5:20 p.m. All participants were required to remain awake during the test day, and all of the tests were given individually in the same experimental setting.

The STAI is a self-reported measure of both state and trait anxiety that comprises 20 items, all rated on a four-point scale, with a higher score indicating greater anxiety. The STAI has demonstrated good validity and internal consistency (Cronbach’s α > 0.85 and test-retest reliability ≥ 0.75) [13]. The SSS is a seven-point self-rating scale used to quantify progressive steps in sleepiness, from 1 (alert) to 7 (no longer fighting sleep).

The computerized WCST, which is loosely considered a measure of frontal lobe ability, consists of four stimulus cards and 128 response cards. Cards differ in the color, form and number of shapes they contain. The examinee is asked to match each response card from the deck with whichever of the four stimulus cards seems most appropriate. The dependent variables included number of perseverative errors, total errors, and categories, as well as percentage of conceptual-level responses and failure to maintain set. Both the DSST and SST of visual attention tasks (VAT) are subsets of the Wechsler Adult Intelligence Scale [12] involving cognitive, perceptual and motor abilities. For the DSST, examinees enter appropriate symbols into empty squares beneath digits. In the SST, they respond to one of two target symbols from four selective symbols. The raw scores of the DSST and SST were determined by the number of items correctly completed in 120 seconds, and the raw scores were then converted to a scale score according to age. The information processing index (IPI) was obtained after transforming the sum of the scale scores of the SST and DSST.

The modified MSLT was performed by partial-montage polysomnography consisting of electroencephalography (at F3/A2, F4/A1, C4/A1, C3/A2, O2/A1, O1/A2), electrooculo-graphy, and submental electromyography. We scored sleep records visually according to the American Academy of Sleep Medicine criteria [14]. The participants lay in a quiet, dark bedroom and attempted to fall asleep. They were awoken after 3 epochs of stage one, or immediately when they entered any other sleep stage. The sleep latency was taken as the first epoch of any stage of sleep. If sleep onset did not occur, a latency of 20 min (the end of the test period) was used for data analysis. The MWT was administered in a similar method as the MSLT. The major differences were that in the MWT, the participants were instructed to remain awake, and the termination criterion was 40 min.

Blood samples were collected by venipuncture (5 ml) in vacuum tubes for sleep-related hormone assays. Serum was separated immediately after blood collection, and the samples were stored at −20°C until analysis. Quantitative determination of cortisol, PRL, GH, and TSH levels in the serum was achieved with a paramagnetic particle, chemiluminescent immunometric assay using a Beckman Access system (Beckman Coulter, Inc., Fullerton, CA, USA), Siemens DPC Immulite 2000 analyzer (Siemens Healthcare Diagnostic Products Ltd. Llanberis, Gwynedd, UK), Siemens DPC Immulite 2000 analyzer (Siemens Healthcare Diagnostic Products Ltd.), and Abbott I2000 (Abbott Ireland Diagnostics Division, Longford, Ireland), respectively. The lower limits of detection were 0.4 μg/dl, 0.16 ng/ml, 0.01 ng/ml, and 0.0025 mIU/L, respectively. The intra-assay coefficient of variation averaged 5% for each item.