Body distribution of impetigo and association with host and pathogen factors

Introduction

Impetigo (also known as skin sores) is a bacterial infection caused by the pathogens Streptococcus pyogenes (Group A Streptococcus or GAS) and Staphylococcus aureus (Carapetis et al., 2005; World Health Organization, 2005). The disease is estimated to affect more than 162 million people worldwide with remote Indigenous communities of Australia regarded as one of the highest areas of disease prevalence (Carapetis et al., 2005; Bowen et al., 2015; May, Bowen & Carapetis, 2016). An estimated 16,000 Indigenous children are affected at any one time (Bowen et al., 2015). Risk factors for developing impetigo are wide ranging and include tropical climate, overcrowding, poverty, scabies co-infection, insect bites and trauma (World Health Organization, 2005).

Skin sores can result in a number of serious complications with high rates of morbidity and mortality (Carapetis et al., 2005; May, Bowen & Carapetis, 2016; McDonald, Currie & Carapetis, 2004; McDonald et al., 2008). These include GAS and S. aureus septicaemia, rheumatic fever with or without rheumatic heart disease, and acute post-streptococcal glomerulonephritis (Carapetis et al., 2005; May, Bowen & Carapetis, 2016; McDonald, Currie & Carapetis, 2004). The adverse health effects also result in substantial socioeconomic consequences. Costly treatments impose a significant economic burden to healthcare systems while absent school days and high transmission rates reinforce the cycle of poverty and low socioeconomic status (World Health Organization, 2005).

Clinical disease generally presents as crusted papules that can arise in bullous or non-bullous forms (Yeoh, Bowen & Carapetis, 2016). GAS has been consistently shown to be the most common causative pathogen in tropical climates while more temperate settings show an increased prevalence of S. aureus (Bowen et al., 2015; Yeoh, Bowen & Carapetis, 2016). There is also growing concern over the increasing rates of community-acquired methicillin resistant S. aureus (CA-MRSA) causing skin sores (Bowen et al., 2015; Tong et al., 2011; Bowen et al., 2014a; Tong et al., 2015). Treatment options include topical antimicrobial agents for limited impetigo and systemic therapy for more extensive disease or in areas of high prevalence (Yeoh, Bowen & Carapetis, 2016; Bowen et al., 2014a).

The broad anatomical distribution of skin sores has been understood as affecting the lower limbs more commonly than the upper limbs and other body regions (scalp, face, neck and torso) (Bowen et al., 2015; Bowen et al., 2014a). However, more detailed descriptions from endemic regions are currently lacking and whether certain disease subgroups are associated with specific skin sore anatomical distribution has not yet been explored.

Materials and Methods

Statistical analysis

We used an inverse distance-weighted interpolation (IDW) (weighting coefficient = 2.0) with a continuous colour gradient from yellow to dark purple to illustrate a low to high density of sores respectively. To ensure standardisation among density distributions, each front and back map for each variable under consideration was assigned the same visualisation parameters including the same colour ramp, a continuous mode and a total of 52 different colour classes.

The range of pixel values ascribed to a given colour output was unique among pairs of front and back maps for each variable of interest. This starts at a lower limit of 0 with an upper limit corresponding to the maximum value of sores on either the front or back map regardless of body region. For example, for the front and back maps of the ‘male’ variable of interest, the range of pixel values was 0 to 80 given that the numerical value of 80 was the highest number of sores for either the front or back map, which corresponded to the ‘right anterior leg’ region (region number 17, Fig. 1).

This was to allow for better colour discrimination among each variable while retaining the same scale for any given front and back map pair while also still allowing for qualitative comparisons between variables. The resultant heat maps were clipped to the boundaries of the template shapefiles and each anatomical region was labelled with the number of sores with totals corresponding to overall trends or depending on the variable under consideration.

Fisher’s Exact Test (FET) with a p-value computed by Monte Carlo simulation using 100,000 replicates was used to compare the distribution of sores between males and females; between sores with only GAS and sores with only S. aureus; and across all ages, with these comparisons chosen a priori. The number of replicates was chosen to ensure that the p-value estimate had stabilized to at least the third decimal place. Statistical analysis was carried out in R version 4.0.5 (R Core Team, 2021).

Results

Stratification by variables

Stratification by sex revealed a similar overall distribution pattern of sores with no observable difference between males and females (Fig. S1; FET, p = 0.154, degrees of freedom (df = 55)). Young children and infants aged 0–2 years had more evenly distributed densities on the posterior surface of the body compared to other age groups (Fig. S2). Children aged 3–6 years start to show a more typical distribution pattern but still have relatively high densities on the upper posterior lower limbs and scalp regions compared to older age groups. Those in the 7–9 and 10–13 years of age groups display a distribution pattern similar to that for overall sores as described above. Overall, differences in the patterns of sore distribution by age were statistically significant (FET, p = 3 × 10⁻⁵, df = 165).

Swabbed sores which cultured both GAS and S. aureus did not show any major observable differences from overall distribution patterns as did those samples which grew GAS only (Fig. 2). Isolates of S. aureus only were found to be more common on the upper posterior lower limbs when comparing to GAS-only sores, with statistically significant differences in the distribution of sores between these groups (FET, p = 3 × 10⁻⁴, df = 55). Skin sores determined to be directly associated with scabies infection predominantly affected the hands (n = 66; Fig. S3), while participants who were noted to have concurrent scabies, regardless of whether associated with their skin sores, showed predominance for the lower limbs (n = 192; Fig. S4). Head lice co-infection predominantly affected the scalp, while those with tinea co-infection showed sporadic distribution although the sample size for each of these variables was low (n = 10 and n = 6 respectively).

Discussion

We have provided a detailed description of the anatomical locations of impetigo lesions in children from an area with endemic rates of impetigo. Exposed areas of the lower and upper limbs were the most commonly affected areas, followed by relatively equal distribution on the face, scalp and upper posterior lower limbs. We found no observable difference between males and females and illustrate that younger children have more evenly dispersed sores on the posterior surface of the body, especially at the lower limbs and scalp.

Overwhelmingly, the pattern of distribution appears to affect more exposed body surfaces, such as the distal lower and upper limbs, as well as the face and scalp when compared to the torso specifically. The increased propensity of impetigo in these areas corresponds with the occurrence of bacterial superinfection after insect bites and trauma (World Health Organization, 2005). Regarding the limbs, these observations would seemingly fit with observed behaviour of children often with uncovered upper and lower limbs due to hot climate. Many of these children then engage in bare-footed play which may result in minor trauma and skin abrasions. Further, scabies co-infection was almost exclusively limited to the hands and corresponds to previously described patterns of scabies infection with resultant pyoderma (Yeoh, Bowen & Carapetis, 2016). This may suggest impetigo of the hands could be largely due to superinfection of burrows of the scabies mite as opposed to scabies-induced pruritus leading to breaks in skin integrity and resultant pyoderma at other body sites.

The call for more targeted preventative measures for impetigo may be further supported by our results. While treatment efforts in high prevalence areas have had substantial impact, encouraged by methods to reduce the cycle of poverty and socioeconomic disadvantage such as improved hygiene practices and appropriate housing, additional strategies are needed to further lower the significant burden of disease that remains (Carapetis et al., 2005; Bowen et al., 2015; May, Bowen & Carapetis, 2016). One method is earlier detection and treatment, and standardised programs such as community skin days and mass drug administration have shown promise (May, Bowen & Carapetis, 2016). Our findings could augment these programs by reminding clinicians to pay particular attention to the body sites most affected, including consideration of a patient’s age group. Furthermore, the use of more protective clothing such as longer upper and lower limb garments as well as encouraging the use of footwear may help reduce skin trauma thereby reducing infection. Impetigo may spread among an individual or via household contacts through cross contamination, and it is not uncommon that those affected have multiple sores (Avire, Whiley & Ross, 2021; Creech, Al-Zubeidi & Fritz, 2015). In this regard, earlier treatment with topical antimicrobial agents could be considered, for example, at body sites with known increased pyoderma distribution which may have a suspicious mild lesion or break in skin integrity after diagnosis at an initial, single body site (Koning et al., 2012; George & Rubin, 2003). However, broader use of topical antimicrobial agents has been associated with the spread of antimicrobial resistance among populations with a high prevalence of skin sores (Williamson et al., 2014). Appropriate surveillance of antimicrobial resistance should be implemented if widespread topical antimicrobials were to be recommended.

Appropriate examination of children for impetigo is important for both opportunistic clinical encounters and for healthy skin screening programs. While exposed body sites in the lower and upper limbs are easily accessible and where most lesions occur, our findings that the buttock and scalp regions are also commonly involved highlight the importance of examining these sites, particularly in younger children. The body maps in our study may be useful for surveillance programs. Changes at the population level in body site distributions of impetigo may indicate shifts in underlying risk factors. For example, an increasing distribution on the hands may indicate an increased prevalence of scabies co-infection. Likewise, an increasing proportion of skin sores distributed in the upper posterior lower limbs may indicate more S. aureus related infections, especially in younger children with nappy-related dermatitis (Ferrazzini et al., 2003; Heath, Desai & Silverberg, 2009). This is in contrast to GAS-only sores which show a distribution pattern similar to overall trends. One reason for this could be that the proportion of GAS-only sores was nearly three times greater than that of S. aureus-only sores and therefore may better reflect the general anatomical distribution. While many risk factors between GAS and S. aureus skin disease overlap such as close household contact, poor hygiene and compromised skin integrity, others such as age of acquisition, seasonal trends and varying geographical prevalence could also account for the minor differences in anatomical distribution we have observed and warrants further exploration (Bowen et al., 2015; Avire, Whiley & Ross, 2021; Bowen et al., 2014b).

Our study has some limitations. The data we have used is limited by the setting, definitions and exclusion criteria of the Skin Sore Trial (Bowen et al., 2014a). This includes limiting the analysis in our study to the two most severe sores if there were multiple sores noted for a participant. While multiple sores (greater or equal to 2) was commonly noted within the trial and may impact our results, our aim was to illustrate the overall trends of impetigo body distribution and allow for a standardised comparison between key variables. Further, compared to older children within the trial, infants may have been more likely to have external clothing removed during examinations, hence resulting in a bias to finding skin sores in the upper posterior lower limbs. Our results may also not accurately reflect skin sore body distribution in other regions of Australia and other countries given data was collected from remote Aboriginal communities of the Northern Territory. The distribution of skin sores in temperate and high-income countries is likely to be different (Bowen et al., 2015). Where impetigo is principally caused by S. aureus, the distribution is also likely to be different. Ectoparasite infections can be difficult to diagnose and both scabies and head lice may have been under-diagnosed. However, the diagnoses were made in the context of a prospective clinical trial with trained staff and the rates of scabies were in keeping with other studies from a similar setting (La Vincente et al., 2009; Andrews et al., 2009).

Conclusions

We have visually represented the distribution of impetigo on the human body in remote Indigenous communities of the Northern Territory, Australia, with stratification by sex, age and causative pathogen. These results highlight key areas of the body for clinicians to pay attention to when examining children for impetigo.

Supplemental Information