A Comparison of Dacron versus Flocked Nylon Swabs for Anal Cytology Specimen Collection

Anal cytology is currently recommended to screen for anal neoplasia among high-risk populations such as men who have sex with men [1]. Yet, anal cytology suffers from low clinical sensitivity for high-grade anal intraepithelial neoplasia, even in human immunodeficiency virus (HIV)-infected people who might have more easily detected, larger lesions [2,3,4]. Dacron swabs are most commonly used for sampling the anus, but it is possible that flocked nylon-tipped swabs may have improved sensitivity because of increased cellular yields. Specifically, flocked nylon swabs have a high surface area and, unlike the woven surface of Dacron swabs, their surface fibers are oriented perpendicularly to the tip so cells are more easily eluted. Although there are no published directed comparisons of the two materials, a previous study of flocked nylon swabs demonstrated an increased sensitivity for cervical detection of human papillomavirus (HPV) and Chlamydia trachomatis DNA compared with rayon-tipped swabs [5], which are quite similar to Dacron.

To examine the best sampling device for collecting anal cells prior to initiating a screening study, we compared the performance of the Dacron and flocked nylon swabs in a pilot sample of 23 HIV-positive men undergoing anal cancer screening by digital rectal examinations and high-resolution anoscopy at Kaiser Permanente San Francisco (San Francisco, Calif., USA). At each examination, 2 anal specimens were sequentially collected from the anal canal. The order of collection was previously assigned by randomization to control for possible differences in quality of cell collection at the first versus second swab. Specifically, the clinician (S.B.) gently inserted the wetted swab into the anal canal until it reached the distal rectal vault (5–6 cm from the anal verge) and then withdrew the swab with rotation and lateral pressure over 15–20 s to ensure the cells were collected from the anal transformation zone. Specimens were placed into Preservcyt medium (Cytyc Corporation, now Hologic, Marlborough, Mass., USA) and irrevocably anonymized with an ID label that was not linked to the patient. All participants provided signed informed consent, and the research was approved by both Kaiser Permanente and National Cancer Institute human subjects review boards.

Specimens were tested for cellularity using a validated molecular assay that quantified a specific genomic DNA sequence (erv-3), a human endogenous retrovirus gene which is present as 2 copies per diploid cell [6,7,8,9]. This allows direct estimation of human cell equivalents per unit volume without a concern of amplification of pseudogenes, which is common for other human gene targets. In addition, slides were prepared, and the number of cells per slide was assessed by digitizing the slides with the Hamamatsu Nanozoomer HT Scan System and applying a previously validated automated algorithm for cell detection [10]. Cellular counts were log-transformed and compared using t tests, and the order effect was taken into account using crossover analysis methods [11]. Although minimum cellularity is often used in the laboratories, no benchmark threshold has been established, and we therefore chose to compare mean values as an indicator of increased or decreased likelihood that specimens will meet a given threshold.

To measure the comparative performance of the swabs for detection of HPV DNA, the specimens were tested by linear array [12,13].

We used 2-sided p values to determine whether flocked swabs performed better or worse than Dacron swabs.

Figure 1 illustrates the log-transformed cell counts based on erv-3 quantification and slide imaging for first and second collections separately, with the means for each strata denoted by larger circles. Table 1 presents the corresponding mean values, standard errors, 95% confidence intervals and 2-sided p values. On average, flocked swabs collected more cells per milliliter of Preservcyt based on erv-3 quantification than Dacron swabs for the first sample, although this was not statistically significant (p = 0.18). For the second sample, Dacron and flocked swabs collected a similar number of cells (p = 0.85). Using cell counts based on slide imaging, the flocked swabs collected on average more cells per slide than the Dacron swabs for both the first and second collections, although not attaining statistical significance for the first sample (p = 0.42 and 0.03 for first and second samples, respectively).

Using standard crossover trial analysis methods for the pooled data to take into account the period effect, we found that flocked swabs clearly outperform the Dacron swabs using cell count per slide based on slide imaging (p = 0.03). However, based on erv-3 quantification per milliliter, although flocked swabs yield a higher number of cells on average than Dacron, they are not statistically significantly different (p = 0.14).

The performance between the swabs for HPV linear array was similar (90.9% agreement, ĸ = 0.804 for 1 of 13 carcinogenic HPV genotypes). No difference in discomfort or bleeding was noted between the sampling devices.

Our study is limited by small sample size and the resulting imprecision of the mean estimates. The data suggest that flocked nylon swabs might have yielded more anal cells than Dacron swabs. Greater cellular yields may increase the representation of diagnostically important cells and reduce the number of anal cytology slides judged as inadequate. Unfortunately, cytopathology reading was not possible for these specimens. The similar performance between swabs for HPV DNA testing suggests that swab type might be more important for cytology sampling compared to HPV sampling. These findings emphasize the need to further investigate the possible clinical utility of using flocked versus Dacron swabs. Although Dacron swabs present a minimal cost, consideration of the increased cost of flocked nylon swabs (≥6 times more expensive) is also required.

Subsequent larger evaluations on optimizing anal cellular sampling are warranted using cytology outcomes as an end point.