Prevalence of macrolide resistance genes present in pharyngeal microbiota
At baseline throat samples were taken in 36 (n = 36/47; 77%) patients in the placebo group and in 44 (n = 44/45; 98%) patients in the azithromycin group. At month 6 and 12 the available samples were, 30 and 27 in placebo group and 34 and 32 in azithromycin group, respectively. The eleven patients in the placebo group and one patient in azithromycin group without a throat sample granted no permission to take an extra throat swab.
Of the 80 patients with a baseline sample taken, 58 (72%) also had a sample at ‘month 6’ and 53 (66%) at ‘month 12’. A total of 43 patients (54%) had samples at both ‘month 6’ and ‘month 12’. This percentage was comparable between treatment arms: 20 (56%) in placebo and 23 (52%) in azithromycin.
The macrolide resistance gene mefA was present in all available throat samples at all time points.
Before treatment, prevalence of the macrolide resistance genes ermF and ermB were respectively 44.4% (n = 16/36) and 86.1% (n = 31/36) in the placebo group (n = 36), and respectively 59.1% (n = 26/44) and 97.7% (n = 43/44) in the azithromycin group (n = 44) (p = 0.261 ermF, p = 0.085 ermB) (Table 3).
Table 3
Prevalence of ermF and ermB macrolide resistance genes over time
| ermF % (pos/all samples) | ermB % (pos/all samples) |
Prevalence | Placebo | Azithromycin | P value | Placebo | Azithromycin | P value |
Baseline | 44,4 (16/36) | 59,1 (26/44) | 0.261 | 86,1 (31/36) | 97,7 (43/44) | 0.085 |
M6 | 43,3 (13/30) | 67,6 (23/34) | 0.050 | 80,0 (24/30) | 97,1 (33/34) | 0.029* |
M12 | 48,1 (13/27) | 68,8 (22/32) | 0.109 | 74,1 (20/27) | 100,0 (32/32) | 0.002* |
* Prevalence of ermB is statistically significant in the Azithromycin group at M6 and M12 compared to the Placebo group (Chi-square, Pearson corrected) |
After 6 and 12 months of placebo treatment, the ermF and ermB genes were detected in 43.3% (n = 13/30), 80% (n = 24/30) at 6 months, and 48.1% (n = 13/27) and 74.1% (20/27) at 12 months of the throat samples tested, correspondingly, with no statistical differences regarding the presence of resistance genes between the treatment groups.
Regarding the azithromycin group, the prevalence of the ermF and ermB genes at 6 months was 67.7% (n = 23/34) and 97.1% (n = 33/34) versus 68.8% (n = 22/32) and 100% (n = 32/32) at 12 months (p = n.s.). Comparison of the ermF prevalence between the placebo and azithromycin groups showed no significant differences at 6 and 12 months (p = 0.05 and p = 0.109). The difference in prevalence of ermB increased significantly over time in the azithromycin group compared to the placebo treated group (p = 0.029 6 months, p = 0.002; 12 months).
Loss and acquisition of macrolide-resistance in pharyngeal microbiota during and after treatment with placebo or azithromycin
In the placebo group, 27 patients had throat swabs available from visits at baseline and 6 months while 26-paired samples were available from baseline and 12 months. For the azithromycin group, there were 34 paired samples (from baseline and 6 months) and 30 pairs (from baseline and 12 months).
The loss and acquisition of macrolide resistance genes (mefA, ermF and ermB) in pharyngeal microbiota before and after treatment of the paired samples is shown in Table 3. During the trial, no differences were detected in the presence of the mefA gene in the pharyngeal microbiota.
For the patients without the macrolide genes ermF and ermB present in their pharyngeal microbiota at baseline (nermF=15 and nermB=4 in placebo, nermF=16 and nermB=1 in azithromycin), no statistical differences were observed in the acquisition rates between the placebo and azithromycin treated groups.
However, from the patients with the macrolide genes ermF and ermB present (nermF=12 and nermB=23in placebo, nermF=18 and nermB=33 in azithromycin) none of the patients treated with azithromycin lost the ermF and ermB gene over time, while for the placebo group, 1 and 3 patients lost the ermF and ermB gene after 6 months, respectively. Moreover, in 5 patients in the placebo group, the ermB gene was lost after 12 months, therefore, the number of patients that lost the gene was statistically significant higher in the placebo group compared to the azithromycin group (p = 0.012).
Relative gene abundances of the macrolide resistant genes during and after treatment with placebo or azithromycin
A large part of the patients in both groups already had detectable levels of macrolide genes at baseline. This enabled us to compare the relative abundance of the genes in throat samples to determine the effect of the treatment on the abundance of these genes. Figure 1 depicts the overall abundance change of a resistance gene (log-transformed).
The relative gene abundance of mefA after 6 months of treatment was substantially higher in the azithromycin v versus placebo group − 0.39, n = 26; p = 0.002) (Fig. 1, Table 4). Determining the overall increase or decrease of the abundance of the ermF gene showed that this gene increased over time after treatment with azithromycin (0.86, n = 18 M6 and 1.04, n = 17 M12) compared to the placebo group (-0.14, n = 11 M6 and 0.15, n = 12 M12) as well, which was only significant at 12 months p = 0.0124 (Fig. 1, Table 4). With regard to the macrolide gene ermB, the relative gene abundance was significantly increased over time in the azithromycin group (0.69, n = 33 M6 and 0.89, n = 30 M12) compared to the placebo group (-0.32, n = 20 M6 and − 0.42, n = 18 M12) after 6 and 12 months of treatment (p = 0.01 and p = 0.001, respectively) (Fig. 1, Table 4).
Table 4
Comparison of mean logs ratios of the different macrolide–resistant genes before, during and after treatment
| | Placebo | Azithromycin | |
| | Mean* | SD | Mean | SD | P value** |
mefA | M6 | -0.22 (n = 27) | 1.00 | 0.51 (n = 34) | 0.47 | 0.0001 |
M12 | -0.39 (n = 26) | 1.21 | 0.33 (n = 30) | 0.68 | 0.002 |
ermF | M6 | -0.14 (n = 11) | 1.35 | 0.86 (n = 18) | 0.99 | 0.0687 |
| M12 | 0.15 (n = 12) | 0.94 | 1.04 (n = 17) | 0.80 | 0.0124 |
ermB | M6 | -0.32 (n = 20) | 1.48 | 0.69 (n = 33) | 0.93 | 0.0116 |
| M12 | -0.42 (n = 18) | 1.33 | 0.89 (n = 30) | 1.05 | 0.0013 |
*mean of the log gene abundance ratio compared to baseline |
**p values for comparison of mean abundance at either 6 or 12 months and baseline, by Wilcoxon ranked sum test. |