Table 2.
Impact of participant characteristics on long-term consistent carriage of a S. aureus spa-type versus intermittent carriage, and never observed carriage versus intermittent carriage, in participants returning ≥12 swabs (∼24 months).
| Factor | (effect in multinomial regression model) | S. aureus long-term consistent spa-type carriers (N = 136): n (row %) or median (IQR) | S. aureus intermittent carriers (N = 192): n (row %) or median (IQR) | Long-term consistent spa-type versus intermittent carriage |
Never observed carriers (n = 88):n (row%) or median (IQR) | Never observed versus intermittent carriage |
||
|---|---|---|---|---|---|---|---|---|
| RR (95% CI) | P value | RR (95% CI) | P value | |||||
| Modal CC | CC30 (spa-CC021) (reference category) | 43 (41%) | 62 (59%) | 1.00 | global 0.13 | N/A | N/A | N/A |
| CC15 (spa-CC084) | 14 (29%) | 34 (71%) | 0.49 (0.22;1.08) | 0.08 | ||||
| spa-CC002 | 9 (33%) | 18 (67%) | 0.72 (0.28;1.88) | 0.50 | ||||
| CC22 (spa-CC005) | 16 (73%) | 6 (27%) | 3.17 (1.09;9.24) | 0.03 | ||||
| spa-CC127 | 6 (33%) | 12 (67%) | 0.58 (0.19;1.75) | 0.33 | ||||
| spa-CC160 | 5 (45%) | 6 (54%) | 0.96 (0.24;3.80) | 0.95 | ||||
| CC8 (spa-CC024) | 3 (30%) | 7 (70%) | 0.73 (0.16;3.35) | 0.69 | ||||
| Other CC (<10 people) | 40 (46%) | 47 (54%) | 1.06 (0.56;1.99) | 0.85 | ||||
| Participant characteristics at recruitment | ||||||||
| Age (years) | (Per 10 years) | 60 (47;68) | 50 (26;63) | 1.05 (0.87;1.27)b | 0.58 | 59 (41;65) | 1.27 (1.01;1.59)c | 0.04 |
| Sex | Female | 58 (27%) | 109 (50%) | 1.00 | 51 (23%) | 1.00 | ||
| Male | 78 (39%) | 83 (42%) | 1.58 (0.96;2.58) | 0.07 | 37 (19%) | 0.83 (0.48;1.43) | 0.50 | |
| Student | No | 127 (36%) | 155 (43%) | 1.00 | 75 (21%) | 1.00 | ||
| Yes | 9 (15%) | 37 (63%) | 0.41 (0.15;1.17) | 0.10 | 13 (22%) | 1.69 (0.60;4.78) | 0.32 | |
| Participant behaviour at recruitment | ||||||||
| Number of other household members | Lives alone | 25 (45%) | 22 (40%) | 1.00 | 8 (15%) | 1.00 | ||
| 1 household member | 71 (36%) | 72 (37%) | (trend, per category higher) | 53 (27%) | (trend, per category higher) | |||
| 2/3 household members | 36 (27%) | 77 (57%) | 0.64 (0.45;0.92) | 0.02 | 21 (16%) | 0.81 (0.54;1.22) | 0.32 | |
| 4 or more/shared | 4 (13%) | 21 (67%) | 6 (19%) | |||||
| Participant behaviour at recruitment | ||||||||
| Days since last outpatient appointmenta | (Per year longer) | 289 (83;1087) | 324 (105;909) | 1.11 (1.03;1.21) | 0.01 | 241 (48;1123) | 1.00 (0.91;1.10) | 0.98 |
| Co-morbidities at recruitment | ||||||||
| Has a long-term illness | No | 46 (25%) | 93 (50%) | 1.00 | 48 (27%) | 1.00 | ||
| Yes | 90 (39%) | 99 (43%) | 1.74 (0.98;3.10) | 0.06 | 40 (17%) | 0.58 (0.31;1.07) | 0.08 | |
| Had MSSA previously | No | 130 (33%) | 179 (46%) | 1.00 | 84 (21%) | 1.00 | ||
| Yes | 6 (26%) | 13 (57%) | 0.62 (0.22;1.77) | 0.37 | 4 (17%) | 0.73 (0.23;2.36) | 0.60 | |
| Antibiotic exposure | ||||||||
| Received anti-staphylococcal antibiotics during follow-up | No | 90 (35%) | 111 (44%) | 1.00 | 54 (21%) | 1.00 | ||
| Yes | 46 (29%) | 81 (50%) | 0.58 (0.35;0.97) | 0.04 | 34 (21%) | 0.87 (0.50;1.50) | 0.61 | |
Note: IQR = inter-quartile range; RR = risk ratio; CI = confidence interval. 9 individuals with missing general practice records excluded from analysis (1 persistent, 6 intermittent, 2 non-carriers). See Supplementary Methods for details of model selection; factors with P < 0.10 shown in bold. Number of household contacts included as a trend term. All effects in the same direction as univariable models. See Supplementary Table 3 for univariable effects of these and other factors considered but not included.
6% of individuals with no outpatient appointment recorded set to 10 years, and all those with last outpatient appointment >10 years ago truncated at 10 years. Median (IQR) calculated in those with previous outpatient appointment.
Effect stronger if number of household members not included in model: HR per 10 years older = 1.16 (95% CI 0.98–1.38) P = 0.09. Effects of other factors unchanged.
Effect stronger if number of household members not included in model: HR per 10 years older = 1.32 (95% CI 1.08–1.63) P = 0.007. Effects of other factors unchanged.