Table 1. Characteristics of included studies.
Where studies reported several datasets, these are specified separately.
| 1st Author | Year | Ref. | nsubjects | ntimepoints | Regression weight, % | Notes |
|---|---|---|---|---|---|---|
| Sothern | 1995 | [17] | 11 | 8 | 0.99 | a |
| Späth-Schwalbe | 1998 | [41] | 16 | 4 | 1.01 | |
| Dugué | 1998 | [42] | 22 | 3 | 1.21 | |
| Vgontzas | 1999 | [24] | 8 | 24 | 1.24 | |
| Ündar | 1999 | [37] | 10 | 7 | 0.84 | |
| Späth-Schwalbe | 2000 | [43] | 18 | 3 | 0.99 | |
| Redwine | 2000 | [44] | 31 | 15 | 3.80 | b |
| Haack | 2000 | [45] | 20 | 9 | 1.90 | |
| Vgontzas | 2000 | [46] | 12; 11 | 2; 2 | 0.54; 0.49 | c |
| Baker | 2001 | [47] | 8 | 6 | 0.62 | |
| Haack | 2001 | [48] | 10 | 10 | 1.00 | |
| Haack | 2002 | [21] | 12 | 25 | 1.99 | |
| Vgontzas | 2002 | [35] | 11 | 24 | 1.71 | |
| Vgontzas | 2003 | [34] | 15; 13 | 48; 48 | 3.29; 2.85 | |
| Alberti | 2003 | [49] | 20 | 2 | 0.49 | |
| Irwin | 2004 | [50] | 15 | 4 | 0.95 | |
| Vgontzas | 2004 | [51] | 25 | 48 | 5.49 | |
| Szczudlik | 2004 | [52] | 17 | 4 | 1.08 | |
| Alesci | 2005 | [33] | 9 | 22 | 1.34 | |
| Mehra | 2006 | [26] | 150; 23; 12; 35; 30 | 2; 2; 2; 2; 2 | 6.72; 1.03; 0.54; 1.56; 1.34 | d |
| Burgos | 2006 | [38] | 11 | 8 | 0.99 | |
| Vgontzas | 2007 | [53] | 20; 20 | 48; 48 | 4.39; 4.39 | e |
| Frey | 2007 | [54] | 19 | 15 | 2.33 | b |
| Lindahl | 2007 | [55] | 14 | 4 | 0.89 | |
| Miles | 2008 | [40] | 51 | 5 | 3.61 | |
| Vgontzas | 2008 | [56] | 15; 13 | 48; 48 | 3.29; 2.85 | |
| Knudsen | 2008 | [39] | 15 | 8 | 1.14 | a, f |
| Benedict | 2009 | [57] | 17 | 15 | 2.09 | |
| Peeling | 2009 | [58] | 8 | 5 | 0.57 | |
| Phillips | 2010 | [59] | 7 | 3; 3 | 0.38; 0.38 | g |
| Rief | 2010 | [60] | 60; 52 | 2; 2 | 2.69; 2.33 | |
| Gill | 2010 | [61] | 14 | 13 | 1.60 | |
| Chennaoui | 2011 | [62] | 12 | 6 | 0.93 | |
| Pledge | 2011 | [63] | 6 | 4 | 0.76 | |
| Crispim | 2012 | [64] | 6; 7; 9 | 7; 7; 7 | 0.50; 0.59; 0.75 | |
| Voderholzer | 2012 | [65] | 16 | 6 | 1.24 | |
| Grigoleit | 2012 | [66] | 10 | 7 | 0.84 | |
| Abedelmalek | 2013 | [67] | 12 | 2 | 0.54 | |
| Pejovic | 2013 | [68] | 30 | 24 | 4.65 | |
| Lekander | 2013 | [69] | 9 | 40 | 1.30 | a |
| Agorastos | 2014 | [36] | 11 | 24 | 1.71 | |
| Kritikou | 2014 | [70] | 18; 21 | 24; 24 | 2.79, 3.26 | |
| Karshikoff | 2015 | [18] | 21 | 4 | 1.31 | a, h |
| Total | 1100 | 789 | 100 |
a: Individual participant data were available. b: Some data were given in time relative to sleep onset or wake-up, and were re-coded using mean chronological time as a best approximation. c: Averaged over 3 consecutive days. d: 358 of 385 participants were included in analyses of IL-6. Final n for each sub-group was not given, and was therefore conservatively coded as the lowest possible n in each sub-group. Error bars were denoted as standard deviation, but were coded as standard errors because they were incredibly small for standard deviations. e: Each dataset was said to have 50% of the total participants (n = 41), and both were conservatively coded as n = 20. f: 15 of 16 participants could be identified in the graph. g: The same 7 participants were included twice with a 10-week interval, yielding two different data sets. h: We have previously published data from this study [71].