Sickle cell anaemia (SCA) is a hereditary disorder of the beta-globulin which affects about 1% of all newborns in Africa [Reference Rees, Williams and Gladwin1]. SCA patients suffer from hyposplenism due to hypoperfusion of the spleen [Reference Khatib, Rabah and Sarnaik2]. This predisposes to infections with encapsulated bacteria such as Streptococcus (Str.) pneumoniae, Haemophilus influenzae or Staphylococcus (S.) aureus [Reference Schaumburg3]. Antibiotic prophylaxis is therefore recommended to reduce mortality and morbidity [Reference Hirst and Owusu-Ofori4]. Despite this selection pressure, we recently reported that resistance rates and capsular types of Str. pneumoniae, H. influenzae and S. aureus did not differ between children with SCA and healthy controls [Reference Schaumburg3]. This might be due to a similar genetic background of these isolates. To test this hypothesis we compared genotypes of S. aureus isolated from this cohort and associated certain lineages with Panton–Valentine leukocidin (PVL) which is highly prevalent in Gabon [Reference Schaumburg5].

Ethical approval was received from the regional ethics committee, Lambaréné (CERIL). Legal guardians of the children provided a signed documented informed consent prior to enrolment.

Children with SCA and controls were recruited in a 1:2 ratio. SCA patients (n = 73) who attended the regular consultation for SCA at the Albert Schweitzer Hospital, Lambaréné, Gabon were enrolled. Healthy controls (n = 143) were subsequently screened for the matching criteria of age (±2 years), residence and sex. If we failed to find children meeting all matching criteria we gave priority to age over residence and residence over sex. For two SCA patients we did not find any matched controls and for one SCA patient only one control was enrolled. Exclusion criteria were (i) signs of any type of infection and (ii) blood transfusion for children with unknown sickle cell state in the past 3 months. The two study groups (SCA patients vs. healthy controls) have been described recently and are comparable in terms of mean age (7·43 vs. 7·44 years), proportion of females (42·5% vs. 47·6%) and residence (urban 4·1 vs. 3·5%, semi-urban 82·2% vs. 81·1%, rural 13·7 vs. 15·4%) [Reference Schaumburg3].

A standardized questionnaire on a patient's history and risk factors for S. aureus carriage [age; frequency of hand washing; passive smoking; crowding and sanitation; parents' education, categorized as ‘primary school’ (école élémentaire), ‘junior high’ (collège) and ‘high school/college’ (lycée/université)] was applied (see Supplementary online material).

Nasal and pharyngeal swabs were cultured onto blood agar plates with an aztreonam disc (Oxoid, UK) for 18–24 h. Colonies were tested for positive catalase reaction and latex agglutination test (Pastorex Staph-Plus; Bio-Rad Laboratories, France). PCR was applied to confirm S. aureus species by nuc detection and to detect the PVL encoding genes (luk S-PV/luk F-PV) [Reference Schaumburg5].

S. aureus protein A typing (spa typing) was performed for each isolate [Reference Mellmann6]. If a participant was colonized at different sites with S. aureus isolates belonging to the same spa type, we only included one isolate in the analysis to exclude the bias of multiple samples. Multilocus sequence typing (MLST) was performed exemplarily for one randomly selected isolate of each spa type [Reference Enright7]; sequence types (ST) and clonal complexes (CC) were assigned in accordance to the S. aureus MLST database (http://saureus.mlst.net).

We compared categorical data using the χ² test or Fisher's exact test and calculated the odds ratio (OR) with the 95% confidence interval (CI). To analyse confounders of risk factors for S. aureus carriage, all categorical variables with a P value <0·2 were used for logistic regression analysis with a stepwise backward elimination. The level of statistical significance was set at P < 0·05. Mean values of continuous variables were compared using Student's t test. All analyses were performed with ‘R’ (http://cran.r-project.org) and the package ‘epicalc’.

The proportion of participants that suffered from various infectious diseases during the 12 months prior to enrolment, according to the personal health record file, was significantly higher in SCA patients than in healthy controls with respect to pneumonia (50·7% vs. 2·2%, OR 46·6, 95% CI 13·3–243·3, P < 0·005), urinary tract infection (6·9% vs. 0·7%, OR 10·2, 95% CI 1·1– 482·1, P < 0·02), and skin and soft tissue infection (12·3% vs. 4·9%, OR 2·7, 95% CI 0·9–9·0, P = 0·48). This was reflected by a higher mean number of days of hospitalization for children with SCA compared to controls (6·2 vs. 0·5 days, P = 0·0005).

The S. aureus carriage rates were similar in SCA patients (n = 34, 46·6%) and controls (n = 67, 46·9%). We merged the two groups to assess risk factors for S. aureus carriage in a multivariate analysis. The age group of children aged <5 years and higher education of a parent (i.e. high school or college degree) were independent risk factors for non-carriage (OR 0·22, 95% CI 0·11–0·42, P = 0·00001 and OR 0·49, 95% CI 0·27–0·90, P = 0·02, respectively). To the best of our knowledge the impact of the parents' educational level on S. aureus colonization in children has not yet been investigated. However, our finding is in line with reports from the USA where lower social status is associated with higher risk for the emergence of methicillin-resistant S. aureus (MRSA) [Reference Witte8]. In our study population, MRSA was only found in one SCA patient (t939, ST45) and one healthy control (t3202, ST88) [Reference Schaumburg3].

Frequent hand washing (>3 times per day) was a strong independent risk factor for S. aureus carriage (OR 14·26, 95% CI 3·94–70·40, P = 0·0002). This is surprising as hand washing is seen as one tool to reduce transmission of S. aureus. Re-acquisition might occur when sharing towels with others immediately after hand washing.

In 26 (35·6%) matched case-controls triplets, both the SCA patient and at least one control were colonized with S. aureus. In nine (12·3%) groups, the same S. aureus spa type was found in the SCA patient and at least one healthy control indicating transmission in the community setting. The spa types which were found in SCA patients and their respective control were t084 (n = 5), t062, t148, t3662 and t6694 (n = 1 each). The distribution of S. aureus genotypes according to spa typing and MLST was balanced in both groups (Table 1). The most frequent ST was ST15 (28·81%, n = 34), followed by ST5 (14·41%, n = 17) and ST152 (8·48%, n = 10). In general, it is assumed that all carrier isolates have the potential to become invasive, whatever the clonal lineage the isolate belongs to [Reference Wertheim9]. However, some STs might be more virulent than others [Reference Schaumburg5]. The distribution of STs in our study is similar to the population structure of S. aureus from asymptomatic Malian carriers [Reference Ruimy10]. However, isolates belonging to ST15 and related STs were only the second most frequent ST in a German study, while other STs were only sporadically detected (ST5) or were not found (ST152) [Reference Monecke11].

Table 1. Distribution of genotypes in Staphylococcus aureus isolates from sickle cell anaemia (SCA) patients and healthy children

CC, Clonal complex; ST, sequence type; OR, odds ratio; CI, confidence interval.

PVL-positive isolates are in bold.

The proportion of PVL was 46·6% (n = 55) and did not differ between groups. The PVL prevalence in Gabon is higher compared to asymptomatic carriers from Europe (0·65%) [Reference Monecke11]. PVL can be associated with skin and soft tissue infection or necrotizing pneumonia and could therefore be a risk for life-threatening infections. PVL-positive isolates (n = 55) were clonal and associated with ST15, ST152, ST5 and ST1 (Table 1).

In conclusion, S. aureus isolates from children with SCA and healthy controls share the same genetic background despite frequent hospital stays and antibiotic treatments in SCA patients. This suggests that results from community-based colonization studies with healthy participants (i.e. resistance studies) could be used to guide treatment and prophylaxis of endogenous infections in children with SCA, despite different selection pressure.