Evaluation of Anatomically Designed Flocked Rectal Swabs for Use with the BioFire FilmArray Gastrointestinal Panel for Detection of Enteric Pathogens in Children Admitted to Hospital with Severe Gastroenteritis

Diagnosing diarrheal disease is difficult in part due to challenges in obtaining and transporting a bulk stool specimen, particularly in resource-limited settings. We compared the performance of flocked rectal swabs to that of traditional bulk stool samples for enteric pathogen detection using the BioFire FilmArray gastrointestinal panel in children admitted to four hospitals in Botswana with community onset severe gastroenteritis.

ABSTRACT

Diagnosing diarrheal disease is difficult in part due to challenges in obtaining and transporting a bulk stool specimen, particularly in resource-limited settings. We compared the performance of flocked rectal swabs to that of traditional bulk stool samples for enteric pathogen detection using the BioFire FilmArray gastrointestinal panel in children admitted to four hospitals in Botswana with community onset severe gastroenteritis. Of the 117-matched flocked rectal swab/stool pairs, we found no significant difference in pathogen detection rates between the flocked rectal swab samples and traditional bulk stool sampling methods for any bacterial (168 versus 167, respectively), viral (94 versus 92, respectively), or protozoan (18 versus 18, respectively) targets. The combination of flocked rectal swab samples with FilmArray testing allows for the rapid diagnosis of infectious gastroenteritis, facilitating a test-and-treat approach for infections that are life-threatening in many resource-limited settings. The culture recovery rates for bacterial pathogens utilizing this approach need to be assessed.

INTRODUCTION

Diarrheal disease remains a leading cause of global childhood morbidity and mortality for children under the age of 5 years, with the majority of deaths occurring in lower-income countries, primarily in sub-Saharan Africa (1). This mortality is not completely attenuated by prompt oral or intravenous hydration therapy and zinc; this is consistent with the fact that two-thirds of diarrhea-related mortalities are observed more than 7 days after presentation to medical attention (2). Given the immense burden of diarrheal disease, some have suggested that specific treatment of infectious pathogens might facilitate patient recovery (3). In addition, pathogen-specific identification of diarrheal disease is also required to rapidly implement effective control measures in the context of outbreak investigations and to inform vaccination programs (4–6).

One of the barriers to diagnosing diarrheal disease is the difficulty inherent in obtaining and transporting a bulk stool specimen, particularly for outpatients and/or in resource-limited settings. There is evidence suggesting that even if enteric specimens are successfully collected, delivery to the laboratory is often greatly delayed (7). Given the rapid results that are provided by current molecular diagnostic assays, any delay in specimen collection becomes a proportionately greater factor influencing overall turnaround time, which may limit any positive impact on patient care. The Biofire FilmArray gastrointestinal panel (GIP) assay (Biofire Diagnostics, Salt Lake City, UT) is a fully automated assay that can detect 22 pathogen targets, including bacteria/bacterial toxins, viruses, and parasites, in just over an hour. Such an assay has allowed for significant reductions in turnaround from sample receipt to result reporting compared to conventional testing approaches (8). However, while this addresses the barriers within the laboratory, it does not impact the impediments to diagnosis that are associated with stool collection (9).

In contrast, rectal swabs can be obtained simply and quickly, even in the outpatient setting, and the vast majority of parents find this type of sampling acceptable (7). Given these advantages, multiple investigators have sought to evaluate rectal swab specimens using different culture, molecular, and antigen testing assays with variable results (7, 10–13). We sought to evaluate anatomically designed flocked rectal swab (FLOQSwabs; Copan Italia, Brescia, Italy) samples for use with the bioMérieux Biofire FilmArray GIP compared to bulk stool collection among children with gastroenteritis in Botswana.

MATERIALS AND METHODS

Children were eligible if they were 2 to 60 months of age, had community-acquired nonbloody acute (defined as <14 days in duration) diarrheal disease with a minimum of three watery stools in a 24-h period, and were admitted to the pediatric medical ward of a participating hospital. The study hospitals were Bamalete Lutheran Hospital in Ramotswa, Deborah Retief Memorial Hospital in Mochudi, Scottish Livingstone Hospital in Molepolole, and Princess Marina Hospital in Gaborone (the largest referral hospital in Botswana). Participants were enrolled prospectively as part of a randomized trial investigating the impact of test-and-treat algorithms and probiotics for the mitigation of diarrheal disease (14). Informed consent was obtained for all participants, and the research protocol was approved by the ethics committees at the Botswana Ministry of Health and Wellness (Health Research Development Committee), the University of Pennsylvania, and McMaster University.

Flocked rectal swab (Copan fecal swab; catalog no. 4C028S; Fig. 1) samples were obtained from each child promptly after enrollment and were immediately eluted in the accompanying 2 ml of modified Cary Blair media. Bulk stool samples were collected at that time as well, if available; otherwise, they were collected as soon as possible after rectal swab collection. Stool samples and flocked rectal swabs were transported to the laboratory in cooler boxes containing ice packs and tested the same day or stored immediately at –80°C and thawed to room temperature prior to testing.

FIG 1.

Rectal swab used in the study.

For bulk stool preparation, 132 mg of the stool sample (∼6.6% volume) was placed in 2 ml of Cary Blair medium. This ratio of stool to Cary Blair medium is recommended by the manufacturer and had previously been evaluated (15). Both samples were vortexed for 30 s, and 200 μl of sample was used to test with the bioMérieux Biofire FilmArray GIP in accordance with the manufacturer’s instructions. To assess differences in yield, McNemar’s test was used for matching swab and stool samples. Pathogen detection in swabs was compared to detection in bulk stool.

RESULTS

A total of 117 matched flocked rectal swab/stool pairs were collected from eligible children. The median age of the participants was 9 months (interquartile range, 6 to 15 months), and 62% were male. Seventeen matched swab/stool pairs had been stored at –80°C from a prior study where collections took place from 26 July 2013 to 19 November 2014. The remaining 100 swabs were collected and tested prospectively between 22 August 2016 and 3 October 2017. Almost all (99%) of the swab/bulk stool pairs were collected on the same day at enrollment (n = 97), and 65% of matched swab/bulk stool pairs were tested on the same day (n = 105). Coinfections were common in this high-burden sample: 74% of the participants harbored more than a single enteropathogen (up to a maximum of five, seen in 8% of the participants), and only 7% of the children had no enteropathogens detected.

Assessing the total of all the pathogen targets detected, the flocked rectal swab samples detected a total of 280 pathogen targets, and the bulk stool samples yielded 277 pathogen targets. The matched swab/stool pair GIP testing results are shown in Tables 1 to 3.

TABLE 1.

Comparison of detection of BioFire FilmArray GIP assay bacterial targets of matched flocked rectal swab/stool pairs (n = 117)

Pathogen targeta	No. of swab/stool pairs				McNemar P
	Both samples positive	Swab (+) only	Stool (+) only	Both samples negative
Shigella/EIEC	16	0	0	101	1.00
Campylobacter	18	2	1	96	1.00
Salmonella	1	0	0	116	1.00
ETEC LT/ST	18	3	2	94	1.00
EPEC	43	5	9	60	0.42
EAEC	44	6	8	59	0.79
V. cholerae	0	3	0	114	0.25
C. difficile toxin A/B	6	3	1	107	0.62

^aEIEC, enteroinvasive E. coli; ETEC, enterotoxigenic E. coli; EPEC, enteropathogenic E. coli; EAEC, enteroaggregative E. coli.

TABLE 2.

Comparison of detection of BioFire FilmArray GIP assay viral targets of matched flocked rectal swab/stool pairs (n = 117)

Pathogen target	No. of swab/stool pairs				McNemar P
	Both samples positive	Swab (+) only	Stool (+) only	Both samples negative
Norovirus GI/GII	10	1	0	106	1.00
Rotavirus A	58	2	1	56	1.00
AdV 40/41	11	1	1	104	1.00
Astrovirus	2	0	0	115	1.00
Sapovirus	8	1	1	107	1.00

TABLE 3.

Comparison of detection of BioFire FilmArray GIP assay parasite targets of matched flocked rectal swab/stool pairs (n = 117)

Pathogen target	No. of swab/stool pairs				McNemar P
	Both samples positive	Swab (+) only	Stool (+) only	Both samples negative
Cryptosporidium	7	0	1	109	1.00
Giardia	10	1	0	106	1.00

There was no significant difference between swab and stool matched sample pairs in the detection of bacterial pathogens (168 versus 167, respectively), viral pathogens (94 versus 92, respectively), and protozoan pathogens (18 versus 18, respectively). Focusing on the pathogens for which antimicrobial treatment is generally recommended (9) in the context of severe gastroenteritis requiring admission to the hospital (Shigella/enteroinvasive Escherichia coli [EIEC], enterotoxigenic E. coli [ETEC], Campylobacter, Cryptosporidium, and Giardia); the flocked swab samples identified 75 pathogens, and the stool samples identified 73 pathogens.

DISCUSSION

The primary finding of our study was that anatomically designed flocked rectal swabs were found to perform similarly to traditional bulk stool specimens across analytes on the BioFire FilmArray GIP assay in a sample of children admitted with severe gastroenteritis in a high-burden setting.

Given these results and the fact that rectal swab sampling is more likely to be successful and implementable across a range of settings, it would seem that this method of acquiring enteric contents will be preferable to bulk stool collection for many care providers. Swabs may provide a more efficient and reliable means of sampling and reducing turnaround time from sample receipt to providing appropriate treatment (7, 14, 16).

The acquisition time of enteric specimens has become more relevant now that rapid enteric testing is possible. One of the barriers to a true point-of-care diagnosis of diarrheal disease is the difficulty and time delays in obtaining and transporting a bulk stool specimen (7). Although bulk stool collection is particularly challenging for outpatients, given the relatively short time these children are in contact with health care professionals, it is also difficult for inpatients. In a large Canadian case series of children presenting with confirmed Shiga toxin-producing E. coli infection, the median time from presentation with illness to stool sample collection was 7 h (17). These issues are compounded in resource-limited settings such as Botswana; in a previous study, we demonstrated that no bulk stool samples were obtained in more than 40% of patients within 2.7 h of attempted collection (7). In Botswana and in many low- and middle-income countries, where morbidity and mortality attributable to acute diarrheal disease are substantial, identification of enteropathogens and targeted antibiotic treatment may confer clinical benefit. However, if bulk stools take hours to acquire and transport, these therapies may be considerably delayed. Our study results suggest that flocked rectal swabs would be a suitable alternative sample type for pediatric patients for molecular detection of enteropathogens using the BioFire FilmArray GIP assay.

The treatable pathogens that were detected included Shigella/EIEC, ETEC, Campylobacter, Cryptosporidium, Giardia, and Vibrio cholerae. Our group and others have found relatively high rates of mortality in children in sub-Saharan Africa presenting with moderate to severe acute gastroenteritis, many of which are found to have these treatable enteropathogens detected in their stools (2, 7, 18).

The primary limitation of this study is that culture recovery using flocked rectal swab specimens was not assessed. For flocked rectal swabs, dilution in Cary Blair medium may impair culture recovery in comparison with bulk stool sample culture recovery. Another limitation of the study was the relatively small number of parasite positive samples included in the study.

Given that the GIP assay only provides qualitative detection, the relative amount of pathogen-specific nucleic acid could not be determined in this study. However, prior evaluations have demonstrated that flocked rectal swab samples have had similar or higher pathogen-specific nucleic acid amounts compared to matched bulk stool samples, particularly for bacterial pathogen targets (7, 19). It remains to be seen how rectal swab samples might perform in comparison to bulk stool samples for direct sequenced-based typing methods.

Conclusions.

Specifically designed flocked rectal swab samples appear to have similar yields for enteric pathogens compared to bulk stool samples when a rapid molecular assay is used. Flocked swab samples would greatly facilitate rapid testing for diarrheal disease, and in the context of severe disease, this combination with the FilmArray GIP assay would greatly facilitate a novel “test and treat” approach for childhood diarrhea.