The prevalence of GBS among pregnant women found in this study (5.7%) was lower compared to what was reported in a similar study conducted in Windhoek, Namibia in 2016, whereby the prevalence of GBS colonisation was reported at 13.6% 20166. Different recruitment centres were included in the current study, which comprise of a different population characteristic from the study by Engelbrecht et al. The prevalence in this current study is lower than the 31.6% found in Zimbabwe in 20007, 30.9% in South Africa in 201217; 9, 37% in South Africa in 20198 and 21.2% in Malawi in 201010.
The findings of this current study are similar to those reported in a Mozambican study, whereby a colonisation rate of 1.8% was obtained11. However, in the Mozambican study swabs were inoculated in Todd-Hewitt broth with only gentamycin (8 µg/ml of agar) and the broth was sub-cultured onto human blood agar. The use of gentamycin alone without nalidixic acid can cause overgrowth of enterobacteriaceae organisms over GBS, hence reducing the growth of GBS. Furthermore, human blood used in preparation of blood agar might have contained inhibitory substances against the growth of GBS11.
GBS colonisation was not associated with age, geographic location, marital status, educational level and employment status. This study revealed that 75.0% of colonised women were aged between 20–39 years and 25.0% were aged 40 years and above. There was however no statistical significant association between GBS colonisation and age (P value 0.209). A similar study conducted at Dr George Mukhari hospital, South Africa found majority of GBS colonised women to be aged between 25 and 29 years, which was also not statistically significant (P value 0.07)9. There was a significant association between maternal age and GBS colonisation in pregnant women screened at the maternity and children hospital in Makkah, Saudi Arabia (P value 0.002)22.
In the current study, 83.3% of colonised women were rural based and 16.7% were urban based. This difference was however not statistically significant (P value 0.708). In a similar study conducted in Dr George Mukhari hospital, South Africa, more of the colonised women were dwelling in the urban areas9. This finding was however also not statistically significant (P value 0.08). It was observed in this study that majority of GBS colonised women were not married (75.0%), compared to married women (25.0%). There was however no significant statistical association between GBS colonisation and marital status (P value 0.921). This agrees with the study conducted at Dr George Mukhari hospital, South Africa, which reported that the majority of colonised women were single, divorced or widowed followed by cohabiting women (P value 0.56)9. In another study conducted in Windhoek by Engelbrecht et al., 2016, more single women were colonised compared to married women (P value 0.315)6. Both findings were not statistically significant.
Furthermore, GBS was more prevalent in women who had no formal secondary education (58.3%), with a decrease in colonisation observed as educational level increased, although this was not statistically significant (P value 0.333). A slight decrease in GBS colonisation with increase in educational level was also observed in the study conducted in Windhoek, although this was also not statistically significant (P value 0.729). Educational level was found to have a significant influence on GBS colonisation in pregnant women in the South African study (Dr George Mukhari hospital), whereby women with tertiary level of education were found to have lower colonisation compared to women with and without secondary education (P value < 0.0001)9.
GBS colonisation was higher in unemployed women than those employed as reported in this current study, with percentages of colonisation being 66.7% and 33.3% respectively. This was however not statistically significant (P value 0.234). On the contrary, a slightly large number of GBS colonised women was found to be employed women as revealed in the study conducted at Dr George Mukhari hospital9, and in Windhoek6. The relationship between employment and GBS colonisation was however not statistically significant in the Windhoek study (P value 0.125)6, while in the South African study it was statistically significant (P value 0.003)9.
Group B streptococcus colonisation did not appear to be linked to obstetric factors such as history of miscarriages and still births, as women with no history of still births and miscarriages were largely colonised (100% and 83.3% respectively). These findings did not have any statistical significance (P value 0.613 and 0.471 respectively). However, GBS colonisation had shown to increase with parity, 50.0% in women with parity exceeding two, 33.3% in women with one or two previous deliveries and 16.7% in women without previous deliveries. This was however not statistically significant (P. value 0.659). In the Saudi Arabian study conducted at Makkah, a significant statistical association was found between parity and GBS colonisation22.
Similarly, GBS colonisation did not appear to be linked to history of miscarriages and still births in the South African study, as the carriage rate was greater in women who had no history of miscarriages and still births compared to the opposite. These were equally not statistically significant findings (P value 0.33 and 0.24 respectively)9. In contrast to the findings of this current study, a study by9, 2016 in South Africa reported that women with multiple deliveries exceeding two were less affected by GBS compared to women who had one to two deliveries and those with zero parity. However, there was no statistical significance (P value 0.24)9.
In the current study, GBS showed sensitivity to penicillin, ampicillin, ceftriaxone, vancomycin, chloramphenicol and linezolid by 100%. This correlates with the findings of the study done in Windhoek, where 100% susceptibility was reported to penicillin, ampicillin, ceftriaxone, vancomycin and linezolid. All isolates were found to be sensitive to clindamycin and erythromycin by 100%. This does not correlate with findings of the Windhoek study, whereby 88.9% sensitivity and 11.1% resistance was reported for erythromycin, while for clindamycin 72.6% sensitivity, 18.8% intermediate sensitivity and 8.5% resistance was reported6.
Since only 18 GBS isolates were obtained from the current study, 100% susceptibility to clindamycin and erythromycin might not be a true reflection of the antimicrobial susceptibility patterns of GBS in the two regions, as a bigger number of isolates might have reflected a different picture. Furthermore, this current study revealed a 100% resistance to tetracycline, while the previous study has revealed 94.9% resistance to tetracycline and 5.1% susceptibility thereof. This differences may be attributable to the differences in the locations of the two studies and also different sample sizes.
The findings of our study are in correlation with another study conducted in Ayder referral hospital and Mekelle health centre, Mekelle, Northern Ethiopia, whereby GBS was reportedly 100% susceptible to penicillin, ampicillin, erythromycin, and vancomycin1. However, in the same study, intermediate GBS susceptibility was reported to chloramphenicol by 42.1% and ceftriaxone by 26.3%, which opposes the findings of our study1.
On the contrary, our study findings are different from the findings of a similar study conducted in Ghandi memorial and Tikur Anbessa, Addis Ababa, Ethiopia, where GBS was found to be 55% susceptible to penicillin, 91% sensitive to ampicillin and all isolates except one were susceptible to erythromycin23. Increased resistance to penicillin in this area was attributable to a wide non-prescription use of penicillin due to weak drug control mechanisms23.
Another study conducted in Gabon, Central Africa to determine capsular type distribution and antimicrobial susceptibility of GBS in pregnant women found that all GBS isolates were susceptible to clindamycin, linezolid, vancomycin and benzyl penicillin which correlates with the findings of our study16. Fourteen isolates were however found intermediary susceptible to erythromycin, which contrasts the findings of our study. However, no inducible clindamycin resistance was found in the Gabon study, which draws a parallel between the two studies16.
In the neighbouring South Africa, a study conducted to assess the antimicrobial resistance of S. agalactiae isolated from pregnant women in Garankuwa found that all strains were 100% susceptible to penicillin, ampicillin and vancomycin, which is similarly demonstrated in our study23. The same study however found some strains to be resistant to erythromycin, clindamycin, tetracycline and chloramphenicol by 21.1%, 17.2%, 94.5% and 24.9% respectively, which distinguishes the two study findings23. In a comparable study conducted in Soweto, South Africa to determine the risk of GBS sepsis in infants exposed to HIV, all isolates were penicillin sensitive, and macrolide resistance was observed in 5.5%25.
Generally, GBS has shown uniform sensitivity to penicillin, ampicillin, vancomycin, ceftriaxone and linezolid across many settings. Resistance to tetracycline was reported at 100% in our study compared to many previous studies, even though they also reported quite high percentages of resistance to tetracycline. Resistance to erythromycin and clindamycin was however not reported in our study, as opposed to other studies. These differences may be attributed to variations in capsular types common in different locations. In additions, fewer GBS isolates were obtained in the current study as compared to other studies, hence the differences in percentages of antimicrobial susceptibility.
None of the isolates phenotypically resistant to tetracycline presented with tet(O) gene, which confers resistance against tetracycline in addition to tet(M) gene13. A related study which was conducted in Windhoek has found tet(M) in 114/117 GBS isolates and of these, 111 were tetracycline resistant as interpreted by the Vitek 26. Based on that study, six of the isolates that possessed tet(M) did not appear to be phenotypically resistant to tetracycline6. This does not concur with our findings, as all isolates presenting with tet(M) were phenotypically resistant to tetracycline in our study. Only tetracycline resistance genes were inspected in our study, as all isolates from this study have only shown phenotypical resistance to tetracycline.
Our study found capsular type Ia to be common among the pregnant women screened. A similar study conducted to determine the antigenic distribution of S. agalactiae isolates from pregnant women at Garankuwa hospital in South Africa revealed that capsular type III was the most common capsular type (29.7%), followed by Ia (25.6%), II (15.6%) IV (8.6%), V (10.9%) and Ib (8.6%)17. These implies that the prominent capsular types associated with this population were III, Ia and II. In the current study, capsular type III was not picked up using PCR. Differences in geographical locations of the two studies may attribute to the differences in capsular types found in the two studies.
Another South African study conducted on pregnant mothers attending prenatal clinics in Soweto concluded that capsular type III was more associated with persistent colonisation throughout the study (29%) than Ia (18%) and V (6%)18. In a Zimbabwean study, the following capsular types were identified as follows: la, lb, II, III and V at 15.7%, 11.6%, 8.3%, 38.8% and 24.0% respectively19, which corresponds with our findings. In Australia, GBS capsular type Ia, II and V were the leading causes of invasive infection26.