INTRODUCTION

Fluids are the most common sample received in cytopathology for diagnosis. The common problem faced by all cytologists is the presence of red blood cells (RBC) that is partially or completely obscuring the morphologic details of the cells in the serous fluid, making it difficult for the cytologist to give an accurate diagnosis.[1] They may be pathological, traumatic, or iatrogenic in origin.[2] They are usually associated with malignancy. The diagnostic performance of the cytological study of the fluid may be attributable to the fact that the cell population present in the sediment is representative of a much larger surface area than that obtained by needle biopsy.[3] Hemorrhagic fluids are processed by a variety of techniques in cytology laboratories. The common goal of each technique is the selection and concentration of an adequate number of tumor cells with intact cell morphology, without losing them during processing. Some of the commonly used reagents and methods are glacial acetic acid (GAA), Carnoy’s fixative (CF), Saponin method, normal saline rehydration technique (NSRT),[4a] cellular fixation, and concentration methods.[4b,5] The present study was undertaken to improve the quality of cytology smears of hemorrhagic samples by hemolysis RBC present in the smear background for better cytological assessment by comparing three different techniques, namely, CF, MCF, and NSRT.

MATERIAL AND METHODS

The present study was a prospective study done for 1 year 6 months from November 2012 to March 2014, in the Department of Pathology in a Tertiary Care Rural Medical College.

All fluids received were subjected to physical and cytopathological examination. Physical characteristics including volume, appearance, and color of the fluid were noted. All hemorrhagic effusions of at least 3 ml volume received are included in the study. All clotted samples and all -non-hemorrhagic effusions were excluded from the study.

Processing

The fluids sent to the laboratory were collected under sterile aseptic conditions by the concerned clinicians. No anticoagulant was used during the time of collection. Hemorrhagic effusions received were taken in a clean tube. The fluid was centrifuged at 2000 rpm for 10 min. A total of six smears were prepared from the sediment thus obtained.

CF was prepared by mixing absolute ethanol, chloroform, and GAA in the proportion of 6:3:1, respectively. Modified CF was prepared by mixing 95% ethanol and GAA in the proportion of 6:1, respectively. Two smears each were placed in CF and modified CF for 5 min. Smears were transferred to 95% ethanol. They were stained with Pap and H & E stains.

NSRT

Two smears were dried in an incubator at 37°C for 5 min. Smears were rehydrated by flooding with normal saline for 30 s. Slides were fixed in 95% ethanol. They were stained with Pap and H&E stains.

These slides were then analyzed for

Efficiency in reducing the number of RBCs

The effects of processing on the nuclear and cytoplasmic staining characteristics

Improvement in cellular yield and hence its diagnostic utility.

Statistical analysis

Statistical analysis was performed using the SPSS statistical software package, version 16. The Chi-square test was used to find out the association of different techniques in the reduction of RBC. A P < 0.05 was considered statistically significant.

RESULTS

A total of 942 samples were received out of which 178 were hemorrhagic. Of these, 96 hemorrhagic samples fulfilled the inclusion criteria and were included in the study.

Effect of reduction of RBCs on smear background

CF and MCF showed very similar results of 30–60% reduction [Figures 1 and 2], whereas more than 60% reduction was seen in 85.40% cases by NSRT method [Figure 3]. The reduction has been found out to be statistically significant with a P- < 0.001 [Table 1].

Figure 1:: (a) Reactive mesothelial cells in hemorrhagic effusion treated with MCF ×40 (Pap) (b) Reactive mesothelial cells in hemorrhagic effusion treated with MCF ×40 (H&E).

Export to PPT

Figure 2:: (a) Reactive mesothelial cells in hemorrhagic effusion treated with CF ×40 (Pap). (b) Reactive mesothelial cells in hemorrhagic effusion treated with CF ×40 (H&E).

Export to PPT

Figure 3:: (a) Reactive mesothelial cells in hemorrhagic effusion treated with NSRT ×40 (Pap). (b) Reactive mesothelial cells of hemorrhagic effusion treated with NSRT ×40 (H&E).

Export to PPT

Table 1:: Effect of reduction of RBCs on smear background.

Technique used >60% reduction 30–60% reduction <30% reduction P-value NSRT 82 12 2 <0.001 MCF 14 70 12 CF 15 69 12 Effect on staining characteristics

SRT showed better staining of cells compared with other fixatives and staining results with both Pap and H&E stains were similar as shown in [Table 2]. Focal changes included intense staining of the cytoplasm and light staining of the nucleus. With saline rehydration technique, stain muck and RBCs were absent with clear background, but the nuclear and cytoplasmic staining was lighter in a few areas which did not affect the diagnosis.

Table 2:: Effect on staining.

Technique Stain No effect Focal changes but not affecting diagnosis No. of cases % No. of cases % CF Pap 47 49 49 51 H&E 45 46.9 51 53.1 MCF Pap 47 49 49 51 H&E 46 47.9 50 52.1 NSRT Pap 54 56.8 42 43.2 H&E 56 58.3 40 41.7 Effect on nuclear chromatin

Nuclei at places showed pale staining compared with other cells. Most of these changes were noted with CF on H&E stain. The techniques used had no significant bearing on the nuclear chromatin in most of the cases as shown in [Table 3]. The nuclear borders or their size were not affected by any of the fixatives. Two cases showed shrinkage which might be because of acetic acid.

Table 3:: Effect on nuclear chromatin.

Technique Stain No Yes No. of cases % No. of cases % CF Pap 93 96.9 3 3.1 H&E 89 92.7 7 7.3 MCF Pap 93 96.9 3 3.1 H&E 93 96.9 3 3.1 SRT Pap 93 96.9 3 3.1 H&E 92 95.8 4 4.2 Effect on cell borders

NSRT showed the effect on cell borders only in few numbers of cases compared with Carnoy’s and Modified CF [Table 4].

Table 4:: Effect on cell borders.

Technique Stain Preserved Disrupted No. of cases % No. of cases % CF Pap 83 86.5 13 13.5 H&E 83 86.5 13 13.5 MCF Pap 87 90.6 9 9.4 H&E 82 85.4 14 14.6 NSRT Pap 87 90.6 9 9.4 H&E 86 89.6 10 10.4

Most of them showed fraying of the cytoplasm which is mainly because of acetic acid; also centrifugation itself could cause such effect on cell borders.

Effect on cytoplasmic staining

Majority of the cases studied showed appropriate staining with all the three techniques [Table 5]. Cytoplasmic under staining was noted in similar proportion in all the three techniques, whereas overstaining in H&E stain was higher with MCF. The presence of acid will enhance eosin staining which was seen in 12 cases on H&E staining. CF has chloroform which acts as clearing agent which helped in better staining compared to modified CF.

Table 5:: Effect on cytoplasmic staining.

Technique Stain Under stained Unaffected Over stained No. of cases % No. of cases % No. of cases % CF Pap 7 7.3 83 86.5 6 6.2 H&E 8 8.3 79 82.3 9 9.4 MCF Pap 6 6.2 84 87.5 6 6.2 H&E 6 6.2 78 81.2 12 12.5 SRT Pap 6 6.2 89 92.7 1 1.1 H&E 7 7.3 87 90.6 2 2.1

Saline rehydration technique showed better cytoplasmic staining where no other chemical is used other than normal saline and alcohol.

Effect on special morphological features

Special features were preserved in all the cases of saline rehydration technique but with CF and MCF the background which had RBCs treated with acetic acid itself caused obscuring of the special features [Table 6].

Table 6:: Effect on special features.

Technique Stain Preserved Obscured No. of cases % No. of cases % CF Pap 94 97.9 2 2.1 H&E 94 97.9 2 2.1 MCF Pap 94 97.9 2 2 H&E 95 99 1 1 SRT Pap 96 100 0 0 H&E 96 100 0 0 DISCUSSION

The relative ease for diagnostic approach and comparatively less painful procedure has made aspiration cytology a diagnostic modality for body effusions to point out the etiology for effusions.[6] The cellular population present in the sediment smear of the fluid is representatively much larger than that obtained by needle biopsy. Therefore, cytology has a greater opportunity to retrieve malignant cells compared to biopsy specimen.[3]

Cytomorphological evaluation of hemorrhagic samples can be very difficult and is often not successful. RBCs may partially or totally obscure the diagnostic cells, making morphological interpretation more difficult.[7] The common goal of each technique is selection and concentration of adequate number of cells with intact cell morphology without losing them during processing. To achieve the best possible smears, precaution should be taken at the level of sample collection, preservation, or immediate processing of the sample. This is most important to prevent the coagulum formation or an anticoagulant can be used.[2]

The most common hemorrhagic effusion received in this study was pleural fluid. Pericardial fluid sample was not received during our study period.

Clear background was seen in 85.4% of cases with NSRT in the present study which correlated with most of the studies. Ng et al. assessed 11 grossly hemorrhagic samples and noted complete lysis in all cases. Immersion of air dried smears in normal saline for 30 s effectively lysed all the RBCs but retained epithelial and mesothelial cells. This resulted in a clean background, optimal for cytological assessment. Most probably rehydration leads to more osmotically active intracellular medium in red cells leading to lysis. Breaking up hemoglobin chains into smaller molecules with drying and rehydration may be a possibility. For nucleated cells an equilibrium is maintained with no major influx of water occur with normal saline rehydration. Cytosol in such cells more readily retains osmotic integrity with drying and rehydration. Hypotonic saline, tap water, and aqueous glycerin were proved to lyse nucleated cells as well as RBCs.[9] Rehydration offers the advantage of cleaner background with more vivid morphological interpretation, especially in malignant cases.[10]

Saline rehydration technique in the present study showed no effect on cells stained with H&E (58.3%) or Pap stain (56.8%). There are no studies to compare the staining effect of H&E and Pap stain but Pap stain after air drying is highly unsatisfactory for cytological assessment,[9] several other studies on cervical cytology have used Pap stain after saline rehydration and reported satisfactory staining features for interpretation. In the present study, Pap stained smears of fluid samples was satisfactory and the nuclear features were also better observed.

Many studies have scored retention of epithelial/mesothelial cells by different methods. Ng et al. observed that with air drying and rehydration technique, cell retention could be increased to 78%. The cell loss with wet fixation appeared more marked in the cytocentrifuged smears (50% loss) as compared with the dried smears (31% loss).[9] Preeti et al. observed with NSRT 86.65% cases scored 4 (same as control smears) followed by Carnoy’s technique 72.02% and by GAA in 56.65%.[11] Shabanam et al. observed retention of cells in 70.5% cases with NSRT followed by Carnoy’s technique 57.8% and by GAA in 50.9%. High cellularity was noted in NSRT 58.9% compared to WF 56.4%.[13] Gupta et al. observed that retention of cells in Pap smears was 80.2% with conventional group and 78.2% with air dried rehydration technique.[10] In the present study, no significant loss of cells was observed in the samples processed from all the three methods [Table 7].

Table 7:: Comparative table for reduction of RBC in the background.

Malvi and Anthony, 2000[1](%) Shabnam, 2013[12](%) Jaiwong et al., 2006[6] (%) Preeti 2011[11](%) Weidmann, 1997[15](%) Kirib, 2006[18](%) Shamsi 2007[16](%) Mirzaie et al., 2006[17](%) Present study 2014 (%) GAA 10 3.9 53.33 CF 20 60.8 82 51.7 15.6 NSRT 93 72.5 96.5 91.33 94 85.4 MCF 14.6 95% ethane 86 0 70.9 Cytorich red 92.5 Cytospin 22.5 Filtration 88

Malavi et al. showed nuclear shrinkage in 76.92% cases with CF and 72% cases with GAA whereas present study only one case showed nuclear shrinkage with CF.

Gupta et al. observed distinct nuclear borders in 78.6% cases by conventional group and 76.1% cases by air dried rehydration group on Pap smears.[10] Distinct nuclear borders in squamous and endocervical cells were, respectively, seen in 88.95%, 79.76% cases in air dried group and 97.09%, and 86.48% in wet fixed group.[6] Nuclear border was distinct in all the Pap smears with ARF and WF.[13] Nuclear details were superior in ARF smears compared to WF smears on gastric mucosal cell smears.[14]

Smears stained with Cytorich red exhibited crisp nuclear details while smears stained with cytospin collection fluid showed smudging with less nuclear details.[7] The chromatin pattern was similar in both ARF and WF.[13] The present study does not show much difference on nuclear features in all the three techniques and Pap stain showed better nuclear details [Table 8]. CF showed 7.3% of cases with hazy nuclear feature in the present study. Smudging of chromatin was seen in 68% of cases with dirty and distracting background.[1] It is an excellent fixative in Pap smears for fixation and reduction of RBC in the background which also helps in identifying squamous cells, columnar cells, inflammatory cells and pathogenic organism for easy diagnosis.[8] GAA showed smudging of chromatin in 73% with dirty and distracting background with decrease in proportion of correct diagnosis.[1]

Table 8:: Comparative table for effect on nuclear chromatin.

Technique Malvi and Anthony, 2000[1](%) Jaiwong, 2006[6](%) Gupta et al., 2003[10](%) Dahlstrom et al., 1999[14](%) Present study 2014 (%) H&E Pap Crisp GAA 26.93 CF 32 92.7 96.9 NSRT 92.86 87.9 71.7 25 95.8 96.9 MCF 96.9 96.9 95% alcohol 96.51 73.6 27 Satisfactory 95% alcohol 63 NSRT 29 zHazy GAA 73.07 CF 68 7.3 3.1 NSRT 7.14 12.20 28.3 33 4.2 3.1 MCF 3.1 3.1 95% alcohol 3.48 26.4 0

Smears treated with NSRT, 11.8% cases showed suboptimal staining and blurring with loss of sharpness of nuclear feature and 3.9% cases were not suitable for assessment.[11] However, smudged chromatin was noted in 7.14% of cases with no nuclear artifacts and morphology was retained in various cavity fluids. However, nuclear artifacts were noted in crisp nuclei by saline rehydration technique.[11]

The present study showed well preserved cell borders in 90.6% by NSRT [Table 9].

Table 9:: Comparative table for cell border preservation.

Technique Jaiwong et al., 2006[6](%) Gupta et al., 2003[10](%) Dahlstrom et al., 1999[14](%) Present study 2014 (%) H&E Pap Preserved CF 86.5 86.5 NSRT 75.58 81.8 20 89.6 90.6 MCF 85.4 90.6 95% alcohol 80.23 83.1 3.7 Satisfactory 95% alcohol 54 NSRT 36 Disrupted CF 13.5 13.5 NSRT 24.41 18.2 31 10.4 9.4 MCF 14.6 9.4 95% alcohol 19.76 16.9 33

The present study showed good staining of cytoplasm by smears treated with NSRT by both staining methods which correspond with other studies [Table 10]. Cytomorphological scoring was done in few studies where cytoplasmic and nuclear features were considered together and scored with Ng et al. scoring system proposed by Ng et al. This showed excellent preservation of cytomorphological features by CF 60.6%, GAA 58.8%, and NSRT 52.9%.[11] Preeti et al. observed excellent cytomorphological preservation using the same scoring system in 77.33% of CF, 68% of GAA, and 46.67% of samples processed by NSRT.[15] Ng et al. also proved that excellent cytomorphological details were seen by wet fixation method compared with rehydration technique.[9] Where the above three studies showed CF as best technique for cytomorphological preservation but the present study and Malvi and Anthony have showed that cytomorphological features are well preserved in smears treated with NSRT.[1,9,11,15]

Table 10:: Comparative results for cytoplasmic staining.

Technique Weidmann, 1997[15](%) Jaiwong et al., 2006[6](%) Gupta et al., 2003[10](%) Present study 2014 (%) H&E Pap Unaffected CF 82.3 86.5 NSRT 98.83 87.8 90.6 92.7 MCF 81.2 87.5 95% alcohol 95.34 79 Cytorichred 97 Cytospin 48 Affected CF 17.7 13.5 NSRT 1.16 12.2 9.4 7.3 MCF 18.7 12.4 95% alcohol 4.65 21 Cytorichred 3 Cytospin 52

Malvi and Antony had difficulty in diagnosis in five malignant untreated hemorrhagic effusion cases. They reported that three cases could be diagnosed by CF, four cases after treatment with GAA, and all five cases after treatment with NSRT.[1] In the present study, out of 96 cases, ten were malignant of which only one was suspected clinically and others were diagnosed on cytological examination of fluids. NSRT helped in confirming these cases with clear background and well preserved, stained cytomorphological features [Figures 4-6]. The present study showed with all three techniques, the microorganism was well preserved and with NSRT special stain PAS positive pseudohyphae was noted [Figure 7]. Immunocytochemistry can be performed on direct cytology smears; however, direct smears will not show the same population of cells in all the smears or there might be absence of cells in the smears. PAP stained smears without destaining or after destaining may be used for immunocytochemistry. Because of numerous technical and quality control issues, it is ideal to have cell-block for immunocytochemistry.[19-21] Smears treated with fixative containing GAA has shown high false positive results on hybrid capture 2 human papillomavirus testing.[17] Ammonium chloride-based erythrocyte lysing reagent, similar to that used for flow cytometry is relatively simpler and inexpensive without compromising the immunoprofile integrity in preparing cell-block of hemorrhagic effusions.[19-21]

Figure 4:: (a) Metastatic adenocarcinoma detected in hemorrhagic effusion treated with MCF ×40 (Pap). (b) Metastatic adenocarcinoma detected in hemorrhagic effusion treated with MCF ×40 (H&E).

Export to PPT

Figure 5:: (a) Metastatic adenocarcinoma detected in hemorrhagic effusion treated with CF ×40 (Pap). (b) Metastatic adenocarcinoma detected in hemorrhagic effusion treated with CF ×40 (H&E).

Export to PPT

Figure 6:: (a) Metastatic adenocarcinoma detected in hemorrhagic effusion treated with NSRT ×40 (Pap). (b) Metastatic adenocarcinoma detected in hemorrhagic effusion treated with NSRT ×40 (H&E).

Export to PPT

Figure 7:: Candida pseudohyphal form detected in hemorrhagic ascitic fluid. PAS stain ×40.

Export to PPT

CONCLUSION

Hemorrhagic fluids are most problematic and cause difficulty in diagnosis. The availability of lysing agents in different combinations has led to several studies of these agents on hemorrhagic samples. The common goal of all the techniques is background free of RBC, retention of epithelial/mesothelial cells and well preserved cytomorphological features. The present study highlights the simple, cost-effective, and user friendly saline rehydration technique in the interpretation of hemorrhagic smears. It also emphasizes that cytomorphological evaluation of smears is highly dependent on the quality of the smears prepared and on the appropriate use of processing technique. Any deviation from the same can lead to difficulties in diagnostic interpretation. In a resource poor countries, saline rehydration can serve as a boon in the processing of hemorrhagic samples especially in the context of screening malignancies by cytological examination. It is a subject of further research in the field of immunocytochemistry and molecular diagnostics whether saline rehydration technique can be useful or hinder the interpretation of these tests.