Correlation of Circadian Rhythms of Heart Rate Variability Indices with Stress, Mood, and Sleep Status in Female Medical Workers with Night Shifts

Figures & data

Figure 1 HRV circadian rhythm analysis. Definition of circadian rhythm parameters (A). The 1440 or 288 data points (black dots) extracted from 1-min (SDNN, RMSSD) or 5-min (LF, HF, LF/HF ratio, and LFnu) epochs of 24-hour ECG signals. The red lines are the circadian rhythm fitting curves of SDNN (B), RMSSD(C) and LF(D), HF(E), LF/HF ratio(F), LFnu(G), heart rate(H) and SD1(I) from one individual, respectively. MESOR, M, is a rhythm-adjusted mean that reflects the overall average of the HRV index; A, amplitude, the height of the maximum oscillation to the midline, and double amplitude represents the extent of rhythmic fluctuation within a cycle; θ, acrophase, is the timing of the zenith occurring in each cycle (the reference time point 12:00 AM was set as the value of 0). The significance of the cosine periodic rhythm curves was tested, indicated by the P value (All P < 0.0001).

Abbreviations: MESOR, Midline Estimating Statistic OF Rhythm; SDNN, standard deviation of all normal intervals; RMSSD, square root of the mean of the sum of the squares of differences between adjacent normal intervals; LF, low-frequency power; HF, high-frequency power; LF/(LF+HF), normalized low-frequency power, namely LFnu.

Table 1 The Median (IQR) Values of Selected Characteristics of the Two Female Groups Stratified by Night Shift Work

Variables	Total (N=257)	No Shift (N=54)	Shift (N=203)	P
Age (years)	33.00 (7.00)	37.00 (8.00)	32.00(6.00)	0.001
BMI (kg/cm^2)	21.00 (3.00)	21.00 (2.00)	21.00(3.00)	0.351
PSS score	29.50(8.00)	28.00(7.00)	30.00(9.00)	0.014
DASS-21 score
Depression	10.00 (10.00)	8.00 (10.00)	10.00 (11.00)	0.086
Anxiety	10.00 (10.00)	8.00 (1.00)	10.00 (10.00)	0.010
Stress	16.00 (12.00)	11.00 (13.00)	16.00 (10.00)	0.001
Sleep quality score	3.00 (1.00)	3.00 (1.00)	3.00 (1.00)	0.195
ESS score	7.00 (5.00)	6.50 (4.00)	7.00 (6.00)	0.005
FAS score	29.00 (10.00)	24.00 (12.00)	30.00 (9.00)	0.006
Heart rate (bpm)	78.27(10.43)	78.46(9.50)	78.03(10.76)	0.231
HRV indices
SDNN_24 h (ms)	155.29(42.04)	141.59(45.30)	157.57(38.96)	0.004
RMSSD_24 h (ms)	52.03(28.62)	52.91(24.74)	51.71(29.18)	0.631
LF_24 h (ms^2)	894.03(819.34)	776.95(884.97)	911.19(803.84)	0.447
HF_24 h (ms^2)	620.62(625.91)	642.23(655.19)	612.45(625.51)	0.877
LF/HF_24 h	1.36(0.67)	1.28(0.68)	1.39(0.67)	0.340
LFnu_24 h	0.65(0.13)	0.66(0.12)	0.65(0.15)	0.441

Notes: Mann–Whitney Wilcoxon test was performed. P < 0.05 was regarded as statistically significant, as indicated by bold.

Abbreviations: BMI, body mass index; PSS, perceived stress scale; DASS-21, depression, anxiety and stress scale-21 items; SHS, subjective happiness scale; ESS, excessive sleep scale; FAS, fatigue scale; HRV, heart rate variability; SDNN, standard deviation of all normal intervals; RMSSD, square root of the mean of the sum of the squares of differences between adjacent normal intervals; LF, low frequency power; HF, high frequency power; LFnu, normalized low frequency power, namely LF/(LF+HF).

Table 2 The Comparison of the Three Circadian Pattern Parameters (M, A, θ) for Each HRV Index Between the Two Female Groups Stratified by Night Shift Work

HRV Index	MESOR			Amplitude			Acrophase (h)
	No Shift	Shift	P	No Shift	Shift	P	No Shift	Shift	P
SDNN (ms)	38.60(13.64)	42.35(15.86)	0.007	3.76(3.46)	6.01(6.40)	0.001	7.15(12.00)	4.80(9.80)	0.252
RMSSD (ms)	29.60(16.60)	33.91(17.17)	0.025	7.45(6.29)	11.18(11.56)	0.010	3.00(4.00)	3.30(2:42)	0.598
LF (ms^2)	462.39(360.23)	560.24(474.21)	0.023	101.50(83.00)	142.93(179.13)	0.002	10.92(10.00)	8.10(11.50)	0.628
HF (ms^2)	213.49(233.65)	294.15(322.50)	0.012	110.05(173.90)	179.63(267.27)	0.009	3.37(3.30)	3.35(2.10)	0.384
LF/HF	0.67(0.09)	0.67(0.12)	0.755	0.12(0.01)	0.11(0.09)	0.769	14.15(2.10)	14.37(2.10)	0.033
LFnu	0.67(0.09)	0.67(0.12)	0.758	0.12(0.10)	0.11(0.09)	0.754	14.16(2.09)	14.58(2.12)	0.034

Notes: Data are presented as median (IQR). Mann–Whitney Wilcoxon test was performed and P < 0.05 was regarded as statistically significant, as indicated by bold.

Abbreviations: HRV, heart rate variability; SDNN, standard deviation of all normal intervals; RMSSD, square root of the mean of the sum of the squares of differences between adjacent normal intervals; LF, low frequency power; HF, high frequency power; LFnu, normalized low frequency power, namely LF/(LF+HF).

Table 3 Association of the Severities of Mental Stress and the Circadian Pattern Parameters (M, A, θ) of Each HRV Index in Night-Shift Female Workers Based on Linear Regression Models^a

		SDNN (ms)		RMSSD (ms)		LF (ms^2)		HF (ms^2)		LF/HF		LFnu
		β	P	β	P	β	P	β	P	β	P	β	P
Perceived stress (PSS)	Log M	0.03	0.049	0.05	0.026	0.09	0.026	0.11	0.019	−0.01	0.101	−0.01	0.101
	Log A	0.03	0.583	0.175	0.175	0.07	0.180	0.11	0.091	0.02	0.497	0.03	0.406
	Log θ	−0.13	0.058^#	0.978	0.978	−0.07	0.325	−0.06	0.249	0.02	0.033	0.01	0.275
Stress (DASS-S)	Log M	0.03	0.153	0.04	0.163	0.06	0.157	0.11	0.055^#	0.09	0.055^#	−0.02	0.055^#
	Log A	0.09	0.095^#	0.09	0.129	0.11	0.070^#	0.11	0.162	0.05	0.189	0.05	0.247
	Log θ	−0.06	0.478	0.02	0.740	−0.14	0.107	0.06	0.350	0.02	0.010	0.02	0.043

Notes: PSS < 30 (Ref) vs PSS ≥ 30; DASSS ≤ 10 (Ref) vs DASSS >10. ^aCovariables included age, BMI, coffee, alcohol, smoke and each separate categorical variable. P < 0.05 was regarded as statistically significant, indicated by bold; P < 0.10 was regarded as suggestive, indicated by #.

Abbreviations: M, MESOR; A, amplitude; θ, acrophrase time; PSS, perceived stress scale; DASS-S, depression, anxiety and stress scale-stress; SDNN, standard deviation of all normal intervals; RMSSD, square root of the mean of the sum of the squares of differences between adjacent normal intervals; LF, low frequency power; HF, high frequency power; LFnu, normalized low frequency power, namely LF/(LF+HF); Log, logarithm (Log 10).

Table 4 Association of the Severities of Anxiety, Depression and the Circadian Patterns Parameters (M, A, θ) of Each HRV Index in Night-Shift Female Workers Based on a Linear Regression Model^a

		SDNN (ms)		RMSSD (ms)		LF (ms^2)		HF (ms^2)		LF/HF		LFnu
		β	P	β	P	β	P	β	P	β	P	β	P
Depression (DASS-D)	Log M	0.02	0.264	0.01	0.648	0.06	0.103	0.04	0.421	−0.00	0.947	0.00	0.947
	Log A	0.06	0.188	0.03	0.557	0.05	0.301	0.01	0.884	0.01	0.741	0.01	0.722
	Log θ	−0.13	0.060^#	−0.04	0.531	−0.15	0.038	−0.04	0.510	0.00	0.878	0.00	0.702
Anxiety (DASS-A)	Log M	0.01	0.571	0.00	0.940	0.05	0.293	0.03	0.573	0.00	0.912	0.00	0.912
	Log A	0.10	0.059^#	0.04	0.437	0.07	0.243	0.04	0.603	0.01	0.766	−0.01	0.885
	Log θ	−0.12	0.111	0.04	0.572	−0.14	0.083	0.00	0.999	0.02	0.028	0.02	0.076^#

Notes: DASSD ≤ 9 (Ref) vs DASSD > 9; DASSA ≤ 6 (Ref) vs DASSA > 6. ^aCovariables included age, BMI, coffee, alcohol, smoke and each separate categorical variable. P < 0.05 was regarded as statistically significant, indicated by bold; P < 0.10 was regarded as suggestive, indicated by #.

Abbreviations: M, MESOR; A, amplitude; θ, acrophrase time; DASS-D, depression, anxiety and stress scale-depression; DASS-A, depression, anxiety and stress scale-anxiety; SDNN, standard deviation of all normal intervals; RMSSD, square root of the mean of the sum of the squares of differences between adjacent normal intervals; LF, low frequency power; HF, high frequency power; LFnu, normalized low frequency power, namely LF/(LF+HF); Log, logarithm (Log 10).

Table 5 Association of the Severities of Sleepiness, Fatigue, Sleep Quality and the Circadian Patterns Parameters (M, A, θ) of Each HRV Indices in Night-Shift Female Workers Based on a Linear Regression Model^a

		SDNN (ms)		RMSSD (ms)		LF (ms^2)		HF (ms^2)		LF/HF		LFnu
		β	P	β	P	β	P	β	P	β	P	β	P
Sleepiness (ESS)	Log M	−0.03	0.122	−0.04	0.116	−0.06	0.168	−0.10	0.057^#	0.00	0.688	−0.01	0.441
	Log A	−0.02	0.717	−0.03	0.569	−0.06	0.318	−0.11	0.130	0.10	0.006	0.10	0.005
	Log θ	0.03	0.757	0.09	0.157	−0.07	0.406	0.07	0.265	0.01	0.141	0.01	0.182
Fatigue (FAS)	Log M	−0.00	0.910	0.00	0.935	−0.00	0.966	0.04	0.609	−0.01	0.302	−0.01	0.302
	Log A	0.10	0.133	0.11	0.101	0.05	0.444	0.09	0.339	0.11	0.020	0.13	0.007
	Log θ	−0.20	0.035	−0.07	0.340	−0.12	0.208	−0.09	0.220	0.01	0.276	0.02	0.246
Sleep quality	Log M	−0.02	0.375	−0.01	0.644	−0.06	0.114	−0.03	0.506	0.00	0.692	0.00	0.692
	Log A	−0.07	0.124	−0.04	0.453	−0.03	0.530	−0.04	0.584	−0.00	0.974	−0.02	0.592
	Log θ	0.18	0.011	−0.00	0.946	0.08	0.308	0.04	0.505	−0.02	0.049	−0.02	0.026

Notes: ESS ≤ 10 (Ref) vs ESS > 10; FAS ≤ 21 (Ref) vs FAS > 21; Sleep quality ≤ 2 (Ref) vs Sleep quality > 2. ^aCovariables included age, BMI, coffee, alcohol, smoke and each separate categorical variable. P < 0.05 was regarded as statistically significant, indicated by bold; P < 0.10 was regarded as suggestive, indicated by #.

Abbreviations: M, MESOR; A, amplitude; θ, acrophrase time; ESS, excessive sleep scale; FAS, fatigue scale; SDNN, standard deviation of all normal intervals; RMSSD, square root of the mean of the sum of the squares of differences between adjacent normal intervals; LF, low frequency power; HF, high frequency power; LFnu, normalized low frequency power, namely LF/(LF+HF); Log, logarithm (Log 10).

Table 6 Association of the Number of Sleep Bouts and the Circadian Patterns Parameters (M, A, θ) of Each HRV Index in Night-Shift Female Workers Based on a Linear Regression Model ^a

		SDNN (ms)		RMSSD (m^2)		LF (ms^2)		HF (ms^2)		LF/HF		LFnu
		β	P	β	P	β	P	β	P	β	P	β	P
Sleep bouts	Log M	−0.00	0.765	0.00	0.914	0.00	0.968	−0.00	0.584	0.00	0.919	0.00	0.919
	Log A	0.00	0.628	0.01	0.337	0.01	0.236	0.00	0.745	0.01	0.017	0.01	0.020
	Log θ	−0.01	0.488	0.01	0.405	−0.01	0.146	−0.00	0.741	0.00	0.700	0.00	0.720

Notes: ^aCovariables included age, BMI, coffee, alcohol, smoke and the number of sleep bouts. P < 0.05 was regarded as statistically significant, indicated by bold.

Abbreviations: M, MESOR; A, amplitude; θ, acrophrase time; SDNN, standard deviation of all normal intervals; RMSSD, square root of the mean of the sum of the squares of differences between adjacent normal intervals; LF, low frequency power; HF, high frequency power; LFnu, normalized low frequency power, namely LF/(LF+HF); Log, logarithm (Log 10).

Table 7 Combined Effects of Daily Mean LF, HF, LF/HF Ratio on the Circadian Pattern Parameters (M, A, θ) of Each HRV Index in Night-Shift Female Workers Based on a Linear Regression Model ^b

		SDNN (ms)		RMSSD (ms)		LF (ms^2)		HF (ms^2)		LF/HF		LFnu
		β^s	P	β^s	P	β^s	P	β^s	P	β^s	P	β^s	P
LF_24 h	Log M	0.70	0.001	0.50	0.001	0.98	0.001	0.45	0.004	0.42	0.004	0.42	0.004
	Log A	0.25	0.157	−0.12	0.469	0.92	0.001	0.17	0.286	−0.38	0.030	−0.45	0.013
	Log θ	0.15	0.394	−0.01	0.964	−0.04	0.830	0.14	0.455	0.17	0.397	−0.04	0.836
HF_24 h	Log M	−0.29	0.083^#	−0.18	0.269	−0.53	0.004	−0.17	0.295	−0.32	0.034	−0.32	0.034
	Log A	−0.00	0.981	0.23	0.203	−0.48	0.011	0.01	0.972	0.26	0.156	0.22	0.234
	Log θ	−0.23	0.224	0.04	0.839	−0.11	0.607	−0.09	0.660	−0.13	0.521	0.07	0.717
LF/HF_ 24 h	Log M	−0.12	0.086^#	−0.30	0.001	0.00	0.995	−0.37	0.001	0.60	0.001	0.60	0.001
	Log A	−0.14	0.076^#	−0.25	0.001	0.04	0.637	−0.40	0.001	−0.21	0.011	−0.28	0.001
	Log θ	0.07	0.387	0.00	0.960	0.02	0.852	0.02	0.812	−0.04	0.623	0.10	0.231

Notes: ^bCovariables included age, BMI, coffee, alcohol and smoke, 24-h averages of LF, HF and the LF/HF ratio. β^s represented standardized coefficients. P < 0.05 was regarded as statistically significant, indicated by bold; P < 0.10 was regarded as suggestive, indicated by #.

Abbreviations: M, MESOR; A, amplitude; θ, acrophrase time; HRV, heart rate variability; SDNN, standard deviation of all RR intervals; RMSSD, square root of the mean of the sum of the squares of differences between adjacent RR intervals; LF, low frequency power; HF, high frequency power; LFnu, normalized low frequency power, namely LF/(LF+HF); Log, logarithm (Log 10).