SEARCH

SEARCH BY CITATION

Keywords:

  • Carrier state;
  • nasal mucosa;
  • persistent carriage;
  • risk factor;
  • screening test;
  • Staphylococcus aureus

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Funding
  9. Transparency Declaration
  10. References

Clin Microbiol Infect 2012; 18: 551–557

Abstract

Persistent Staphylococcus aureus nasal carriers are at high risk of S. aureus infection. The present study delineates a simple strategy aimed at identifying rapidly and accurately this subset of subjects for clinical or epidemiological purposes. Ninety healthy volunteers were each identified as persistent, intermittent or non-nasal carriers of S. aureus by using seven specimens sampled over a 5-week period. By reference to this so-called reference standard, six other strategies aimed at simplifying and speeding the identification of persistent carriers and based on the qualitative or quantitative detection of S. aureus in one to three nasal samples were evaluated by the measure of the area under the curve of receiver operating characteristic diagrams. Among strategies using qualitative results, there was no statistical difference between protocols using seven and three samples. A threshold of 103 CFU of S. aureus per swab was found capable of defining persistent nasal carriage with a sensitivity of 83.1% and a specificity of 95.6%. These figures reached 95.5% and 94.9%, respectively, by using an algorithm including one or two nasal specimens according to the threshold of 103 CFU of S. aureus in the first swab. The latter two strategies were shown to be costly equivalents. The proposed algorithm-based strategy proved to be relevant to identify properly and consistently persistent nasal carriers of S. aureus. However, as it was built from data of healthy volunteers, it needs to be confirmed prospectively on patients potentially at risk for S. aureus infection.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Funding
  9. Transparency Declaration
  10. References

The vestibulum nasi is considered the main site of Staphylococcus aureus colonization in humans (for a recent review see ref. [1]). Previous studies identified three patterns including persistent, intermittent and non-carrier states in approximately 20%, 30% and 50% of the whole population, respectively [1]. Persistent nasal carriers of S. aureus can be distinguished from intermittent carriers by a lower exchange rate of S. aureus strains or clones in repeated cultures [2–4] and by a higher mean number of CFU of S. aureus in the nose [1,4].

Patients with persistent carriage have been shown to exhibit a higher risk of S. aureus infection than patients with other statuses [5–7], at least in part by inducing a higher dispersion of S. aureus in the environment [1,4,5]. It has been shown recently that a decolonization procedure using mupirocin and chlorhexidine before surgery enabled a significant rate decrease of S. aureus infection [8]. Targeting patients who would benefit from such a drug-based prevention protocol could minimize the risk of emergence of S. aureus resistance [9,10]. Consequently, simple though fully reliable screening methods are needed to rapidly identify persistent nasal carriers.

However, the characterization of persistent carriage varies from one study to another and there is no consensual definition of the status of S. aureus persistent carrier. In previous studies, this status was determined by using from 5 to 12 consecutive specimens taken over several weeks to months with a positive rate of at least 80% [4,6,11,12]. Seven successive nasal swab cultures was shown to reliably distinguish non-carriers from intermittent carriers [1,13]. Only one study proposed a ‘culture rule’ combining qualitative and quantitative results from only two nasal swabs taken within a 1-week interval to predict the persistent S. aureus carriage state with a positive predictive value of 79% [13].

The present study involving healthy volunteers was designed to simplify the accurate identification of S. aureus persistent carriers. By comparison to a reference protocol using seven successive specimens, a simplified algorithm using quantitative data from one or two nasal swabs in a short interval of time was found to be suitable to accurately segregate persistent carriers from subjects exhibiting a different status.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Funding
  9. Transparency Declaration
  10. References

Study population

Ninety-three volunteers from the University Hospital of Saint-Etienne, France, were included in the study from March to April 2010. Volunteers were healthy healthcare workers, aged 18–65 years, and exhibited none of the following exclusion criteria: acute infection, chronic skin disease, current use of antibiotics and pregnancy. A written informed consent was obtained from all the volunteers. The study received the approval of the regional ethical research committee (‘Comité de Protection des Personnes Sud-Est I’).

Screening

Seven sampling episodes were scheduled to screen S. aureus nasal carriage over a 5-week period. During sampling episodes, nostrils were sampled independently with nylon flocked swabs (Regular nylon flocked swab 552C, microRheologics, Brescia, Italy) and rayon swabs (Fastidious anaerobe swab 108C, Copan, Italy) randomized according to the nostril side [14]. The sampling episodes were scheduled every 2 days for the first three specimens (D0, D2, D4) and once a week for the last four specimens (D7, D15, D23, D31). Samples were taken by one trained researcher following a pre-defined protocol; swabs were wetted in physiological serum, inserted approximately 2 cm into the anterior nostril and rotated five times [14].

Microbiological procedures

Samples were processed within 2 h. Swabs were vortexed for 10 s in 1 mL phosphate-buffered saline and 50 μL of this solution was plated onto an appropriate chromogenic medium (BBL™ CHROMagar™ Staph aureus, Becton Dickinson, Le Pont de Clair, France). Plates were read after 24 h and 48 h of growth at 36°C and pink colonies were plated onto blood agar for identification by matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) (Bruker Daltonics, Bremen, Germany) [15–17]. The bacterial load, expressed as CFU/swab, was determined by the number of pink colonies after 48 h of growth on the chromogenic medium. The limit of detection was 20 CFU/swab of S. aureus and adapted dilutions were performed to quantify high bacterial loads. As the two nostrils were screened independently and in a randomized way, the highest bacterial load was selected for analysis.

Gentoyping of S. aureus isolates

The strains of S. aureus were characterized by arbitrarily primed PCR (AP-PCR) using the primers ERIC2 and 1 referred to in ref. [18], as previously described with minor changes [19]. Briefly, the DNA extraction was performed in a QIAsymphony® machine using a virus/bacteria mini kit (Qiagen, Courtaboeuf, France). The amplification step was performed in a GeneAmp® PCR System 9700 thermocycler (Applied Biosystems, Villebon-sur-Yvette, France). The amplicons were stained with GelRed™ (Interchim, Montluçon, France) and separated in a 1.5% agarose gel (UltraPure™ Agarose, Invitrogen, Cergy Pontoise, France). The AP-PCR profiles were analysed with Quantity One Software (Bio-Rad, Marnes-la-Coquette, France). Strains were considered different if the profiles exhibited more than two bands of difference with at least one of two primers [19]. The susceptibility of each clone to methicillin was studied by detection of the mecA gene.

Definition of nasal carrier state

Assuming that persistent carriers are defined by a positive rate of nasal swabs of at least 80% [6,11,12,20], persons with six or seven cultures positive for S. aureus were classified as persistent carriers; those with negative results of all cultures were classified as non-carriers. All other persons were classified as intermittent carriers. This strategy was labelled the ‘reference protocol’.

Evaluation of screening strategies of persistent S. aureus nasal carriage

Based on previously published protocols looking for nasal carriage of S. aureus [8,13,21–23], six different screening strategies, lettered from A to F, were evaluated. The first three strategies were based only on qualitative results of one (A), two (B) or three (C) consecutive positive samples. Strategies D and E relied on a single sample with a bacterial load of >102 and >103 CFU/swab, respectively. Strategy F was an algorithm combining quantitative results from one or two nasal samples (see Results section for more details).

Cost evaluation

The cost generated by the different screening strategies was calculated using the following items: 1.11 € per swab, 1.34 € per chromogenic medium and 2.44 € per bacterial identification by MALDI-TOF MS technology [16].

Statistical analysis

Assuming a prevalence of persistent S. aureus nasal carriers of 24% [11] and a sensitivity of 95% with a precision of 10%, a number of 73 volunteers was found to be necessary for detecting 18 persistent carriers. The Wilcoxon non-parametric test was used for mean comparison. Predictive values were corrected by reference to a prevalence of persistent nasal carriage of 24% [11], using the Bayes theorem [24]. The SPSS software (version 16.0, Chicago, IL, USA) was used for statistical analyses and calculation of receiver operating characteristic (ROC) curves. The area under the curve (AUC) of the ROC curve of each strategy was compared with that of the reference protocol and p values were obtained following the method of Hanley and McNeil [25] using the ROC curve’s standard error calculated under non-parametric assumption. Values of p below the 5% level were considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Funding
  9. Transparency Declaration
  10. References

Study population

From the 93 volunteers incorporated into the study, three subjects were excluded, respectively, for an antibiotic treatment at D4, a diagnosis of pregnancy at D23, and missing samples. The analysis was performed on 90 subjects comprising 54 women and 36 men (sex ratio of 0.67) with a mean age of 36.5 years (range 19.1–62.7).

Nasal carriage

For each volunteer, seven sampling episodes were performed on a mean period of 31 days (range 25–41). According to the reference protocol described above, 22 (24.4%) volunteers were classified as persistent carriers, 13 (14.5%) as intermittent carriers and 55 (61.1%) as non-carriers. Table 1 illustrates the quantitative results of S. aureus nasal screening in persistent and intermittent carriers. The mean of positive samples per sampling episode was 25.3 (28.1%) with a standard deviation (SD) of 1.48 (21.6 ± 0.49 and 4.0 ± 1.07 for persistent and intermittent carriers, respectively). The amount of S. aureus in nostrils of persistent carriers (median of 3.03 × 105 CFU/swab) was higher than that in intermittent carriers (median of 40 CFU/swab) (p <0.001 by Wilcoxon test) (Fig. 1). All S. aureus isolates (n = 300 from 35 carriers) were genotyped by AP-PCR; 38 different profiles were obtained according to the criteria described above (data not shown). A different clone was identified for at least each subject, except for a couple of persistent carriers whose strains exhibited similar profiles. Of the 22 persistent carriers, 20 harboured a single clone; in the two remaining cases, either two or three clones were isolated but a dominant clone was present in >80% of the samples, which complies with the above definition of persistent carriers. None of the 38 clones was found positive for the presence of the mecA gene.

Table 1.   Nasal bacterial loads of Staphylococcus aureus of persistent and intermittent nasal carriers of the studied cohort
Volunteer no.Nasal S. aureus loads in CFU/swab according to the sequence of sampling episodes (S)Nasal carriage statusa
S1S2S3S4S5S6S7
  1. aValues in parentheses represent number of positive samples/total samples.

  2. bSamples not performed.

0033.8 × 1021.5 × 1042.0 × 1011.1 × 1051.0 × 107Intermittent(5/7)
0042.0 × 1012.0 × 101Intermittent(2/7)
0073.4 × 1022.6 × 1025.5 × 1052.9 × 1032.5 × 1033.5 × 105Persistent(6/7)
0091.6 × 1041.7 × 1042.2 × 1063.4 × 1051.8 × 1041.6 × 1072.6 × 106Persistent(7/7)
0112.2 × 1032.5 × 1042.3 × 1041.4 × 1052.5 × 1063.5 × 1052.3 × 103Persistent(7/7)
0142.0 × 101Intermittent(1/7)
0152.5 × 1036.4 × 1038.0 × 1021.2 × 1042.6 × 1021.7 × 1033.0 × 102Persistent(7/7)
0181.8 × 1022.1 × 1031.0 × 1021.3 × 1063.1 × 1052.4 × 1069.9 × 103Persistent(7/7)
0259.5 × 1051.3 × 1043.6 × 1055.0 × 1041.9 × 1063.9 × 1053.7 × 105Persistent(7/7)
0274.0 × 101Intermittent(1/7)
0281.0 × 1024.0 × 1011.4 × 1021.6 × 1022.0 × 101Intermittent(5/7)
0292.0 × 101Intermittent(1/7)
0342.0 × 101Intermittent(1/7)
0362.3 × 1051.3 × 1046.2 × 1056.9 × 1057.5 × 1051.1 × 1042.7 × 106Persistent(7/7)
0373.8 × 1063.2 × 1045.0 × 1043.6 × 1056.3 × 1064.3 × 1045.5 × 106Persistent(7/7)
0394.0 × 101Intermittent(1/7)
0404.0 × 1014.0 × 1019.4 × 1022.9 × 1039.0 × 102Intermittent(5/7)
0474.2 × 1039.8 × 1052.3 × 1031.1 × 1042.1 × 1054.2 × 1057.3 × 104Persistent(7/7)
0505.8 × 1061.9 × 1052.2 × 1061.1 × 1072.2 × 1025.0 × 1022.9 × 106Persistent(7/7)
0517.3 × 1057.0 × 1042.9 × 1061.5 × 1073.9 × 1066.9 × 107bPersistent(6/6)
0521.2 × 1023.6 × 1023.8 × 1027.2 × 1023.4 × 1026.6 × 102Persistent(6/7)
0581.5 × 1052.9 × 1066.1 × 1053.0 × 1024.1 × 1064.6 × 1072.5 × 106Persistent(7/7)
0595.7 × 1068.7 × 1051.8 × 1052.2 × 1053.3 × 1061.3 × 1064.8 × 106Persistent(7/7)
0611.4 × 1062.2 × 1067.2 × 1021.1 × 1034.4 × 1054.5 × 1059.0 × 105Persistent(7/7)
0672.0 × 1051.2 × 1031.5 × 1061.2 × 1072.6 × 1052.0 × 1013.6 × 102Persistent(7/7)
0693.2 × 102Intermittent(1/7)
0702.1 × 1063.1 × 1061.8 × 1053.2 × 1061.4 × 1063.7 × 1063.2 × 107Persistent(7/7)
0712.8 × 1051.4 × 1081.5 × 1071.5 × 1072.0 × 1053.3 × 1063.1 × 105Persistent(7/7)
0732.0 × 101Intermittent(1/7)
0762.0 × 1012.0 × 101Intermittent(2/7)
0792.0 × 1032.0 × 1034.0 × 1024.6 × 1021.1 × 1031.2 × 1022.3 × 103Persistent(7/7)
0804.2 × 102Intermittent(1/7)
0823.0 × 1051.4 × 1063.6 × 1078.7 × 1042.0 × 1058.1 × 1043.0 × 104Persistent(7/7)
0872.8 × 1078.0 × 1051.0 × 1062.3 × 1064.7 × 1053.0 × 1041.6 × 106Persistent(7/7)
0901.8 × 1062.2 × 1069.5 × 1051.0 × 1042.1 × 1066.4 × 103bPersistent(6/6)
image

Figure 1.  Comparison of bacterial load according to the nasal carriage status (p <0.001 by Wilcoxon test).

Download figure to PowerPoint

Screening strategies for identification of persistent S. aureus nasal carriers

Based on qualitative and quantitative data reported in Table 1 and Fig. 1, different screening strategies aimed to distinguish between persistent and intermittent carriers were evaluated by AUC of ROC curves in comparison to the AUC of the reference protocol. As shown in Table 2, strategy C based on three positive consecutive samples provided the highest performances among strategies using qualitative data; the AUC being not statistically different from that of the reference protocol using seven sampling episodes. When the data in Table 2 were re-analysed by comparing persistent carriers with other subjects (including non-carriers and intermittent carriers), all the strategies were found to be not different from the reference protocol, with the exception of strategy E, which remained less efficient (data not shown).

Table 2.   Description of different screening strategies proposed to identify Staphylococcus aureus persistent carriers in comparison to the reference protocol (consisting of at least six positive out of seven consecutive samples)
No.Screening testSensitivity, %Specificity, %PPV, %NPV, %AUCp value*Cost in €a
  1. PPV, positive predictive value; NPV, negative predictive value; AUC, area under curve.

  2. Values in parenthesis represent the 95% confidence interval. Data of non-carriers (defined by no positive sample) were excluded of the data analysis.

  3. *p value was calculated by comparison of the AUC value reported in the previous column to the AUC value of the reference protocol (arbitrarily fixed to 1) following the method of Hanley and McNeil [25].

  4. aValue in parenthesis represents the cost adding identification by MALDI-TOF MS technique performed in positive sample (28%).

  5. bThis strategy is detailed in Fig. 2.

A1 positive sample97.4 (80.5–99.8)70.3 (40.0–91.1)52.5 (26.2–77.4)98.8 (79.7–100)0.84 (0.68–0.94)0.012.45 (3.14)
B2 positive consecutive samples95.5 (77.4–99.8)88.5 (60.0–98.2)75.5 (42.2–92.9)98.3 (83.3–100)0.92 (0.78–0.98)0.074.90 (6.27)
C3 positive consecutive samples93.6 (74.6–99.5)95.4 (68.5–99.9)87.8 (50.5–98.5)97.9 (82.7–100)0.95 (0.81–0.99)0.147.35 (9.41)
D1 sample >102 CFU/swab96.1 (78.4–99.8)87.9 (59.5–98.0)75.4 (42.4–91.8)98.5 (82.5–100)0.92 (0.78–0.98)0.052.45 (3.14)
E1 sample >103 CFU/swab83.1 (61.5–95.2)95.6 (69.1–99.7)88.3 (48.3–97.6)94.6 (78.2–99.6)0.89 (0.75–0.97)0.042.45 (3.14)
F1 or 2 quantitative samplesb95.5 (77.4–99.8)94.9 (68.1–99.6)87.7 (51.7–97.6)98.5 (83.5–100)0.95 (0.82–0.99)0.122.62 (3.36)

By combining quantitative results, a new algorithm (strategy F) using one or two sampling episodes was built (Fig. 2): a carrier was classified as persistent if the first sample recovered >103 CFU/swab or, for counts >102 CFU/swab and ≤103 CFU/swab, if the second sample recovered >102 CFU/swab. As for strategy C, strategy F was shown to exhibit performances that were not different from that of the reference protocol (Table 2). Using this algorithm, 92% of the cohort volunteers were segregated as persistent carrier (>103 CFU/swab) or non-persistent carrier (≤102 CFU/swab) by using a single sample whereas a second sample was needed to determine the carriage status of the remaining volunteers. The costs of the different strategies are shown in Table 2: the cost of strategy F was very close to that of strategies using a single swab.

image

Figure 2.  Definition of the persistent carriage state by the proposed algorithm in the studied cohort.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Funding
  9. Transparency Declaration
  10. References

Staphylococcus aureus nasal carriage is the main risk factor for surgical site infection [1,26–29] and among S. aureus nasal carriers, persistent carriers have the highest risk of infection with this bacterium [1,6,11]. As recently illustrated [8], preventive decolonization of S. aureus carriers reduces the risk of surgical site infection. However, the systematic use of antibacterial agents to decolonize patients showed a dramatic, reciprocal increase of mupirocin [9] or chlorhexidine [10] resistance. Furthermore, high rates of recolonization were observed after mupirocin ointment, mainly with the endogenous strain, suggesting the persistence of S. aureus in niches potentially associated with persistent carriage status [1,30–32]. According to the recent reclassification of nasal carriers proposed by van Belkum et al. [11], it might be worthwhile to reconsider decolonization strategy according to the actual carriage status. There is an urgent need for rapid microbiological tools able to segregate accurately persistent nasal carriers of S. aureus from other groups and to propose a targeted preventive decolonization to patients with a high risk for S. aureus infection. This study reports a new algorithm that, in a cohort of healthy volunteers, was shown able to meet this need by using one or two sampling episodes together with a quantitative culture method.

Despite the absence of a consensual definition of nasal carriage states of S. aureus and assuming that a persistent carrier could be defined by using seven samples taken during a 5-week period, the screening of S. aureus nasal carriage reported in this study exhibited prevalences of 24.4%, 14.5% and 61.1% for persistent carriers, intermittent carriers and non-carriers, respectively. These rates are in accordance with those previously described [1,11], even if the prevalence of intermittent carriers is lower than expected. The probability to identify a volunteer as an intermittent carrier increases with the number of sampling episodes and the duration of the screening period, as shown by the higher prevalence of intermittent carriage in studies performed over a long time [3,6,11,12]. A possible under-estimation of intermittent carriers is not problematic in the present study because its objective was mainly to distinguish properly persistent carriers from the other subjects. Accordingly, a 5-week period of screening using seven sampling episodes was considered sufficient to constitute an accurate reference standard for testing new strategies aimed at identifying persistent nasal carriers of S. aureus.

The main limitation of this study is the absence of validation of the algorithm described in Fig. 2 on a cohort of patients at risk for S. aureus infection, as was performed by Nouwen et al. [13] for the validation of their ‘culture rule’ using a panel of elderly persons from the Rotterdam cohort. This step is in progress in our hospital. Another limit of the study is the fact that the new algorithm requires an excellent repeatability of the sampling method to quantify properly the bacterial load. The above results were performed with samples taken by one researcher trained in the nasal sampling technique. In routine practice, these samplings are performed by heterogeneous personnel, which could affect the reliability of the algorithm. As discussed by Nouwen et al. [13], the training of healthcare workers to perform sampling methods is essential to the success of strategies relying on quantitative microbiological techniques. Moreover, considering the variability of nasal bacterial loads of S. aureus in different clinical settings [13,23], the accuracy of the new algorithm should be verified prospectively in these specific populations.

To verify that the population of healthcare workers tested in this study was not biased in terms of S. aureus nasal carriage, we showed that (i) all but two colonized subjects harboured non-clonal strains and (ii) subjects defined as persistent carriers were indeed colonized by the same strain of S. aureus (even if, for some volunteers, another strain was also present in a few samples). The latter finding allows the exclusion of the misclassification of intermittent carriers successively colonized by varying clonal lineages by daily contact with infected or colonized patients as persistent carriers. The absence in our cohort of S. aureus strains resistant to methicillin provides more evidence that the isolated bacteria were part of the nasal endogenous flora of the volunteers and not inherited from nosocomial transmission.

Nevertheless, according to the algorithm, >90% of the volunteers were classified properly after the first nasal sampling episode. Assuming that these results could be extrapolated to surgical patients, the patients scheduled for surgery could be screened at the time of the pre-surgery visit and a minority of them would need to be sampled again at the anaesthesia visit or at hospital admission. So, in the majority of cases of planned surgery, the preventive decolonization could be scheduled and started before the hospitalization of patients.

To simplify and accelerate the microbiological procedures, flocked swabs and chromogenic media were used together with identification by MALDI-TOF MS technology. The technique could still be simplified by using a nylon flocked swab suspended in Amies liquid transport medium, which would allow direct plating of 100 μL transport medium onto the appropriate chromogenic medium without the need for any additional dilution.

Strategies using a real-time PCR assay with [33,34] or without an integrated extraction step [8,35] are able to detect the presence of S. aureus colonization within a few hours. However, it cannot currently help in predicting the carrier state and is expensive by comparison with standard culture. In contrast, the strategy proposed in this study is much cheaper and could provide a result within 24 h, with an accurate indication of the carrier state in most cases.

In conclusion, an algorithm based on one or two nasal sampling episodes using flocked nylon swabs coupled to chromogenic culture was found able to identify properly and at low cost nasal carriers of S. aureus in a cohort of healthy volunteers. The following step will comprise the validation of this algorithm in various clinical settings and under less standardized sampling and microbiological conditions.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Funding
  9. Transparency Declaration
  10. References

All the volunteers are acknowledged for their free participation in the study. Paul Verhoeven thanks Maria Rodrigues for her skilful technical assistance.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Funding
  9. Transparency Declaration
  10. References

This work was supported by a grant from the University Hospital of Saint-Etienne, France.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Funding
  9. Transparency Declaration
  10. References