Article Text

Download PDFPDF

Increased lymph node harvest from colorectal cancer resections using GEWF solution: a randomised study
  1. L H Iversen1,
  2. S Laurberg1,
  3. R Hagemann-Madsen2,
  4. H Dybdahl2
  1. 1Department of Surgery P, Aarhus University Hospital THG, Aarhus, Denmark
  2. 2Department of Pathology, Aarhus University Hospital THG, Aarhus, Denmark
  1. Dr L H Iversen, Department of Surgery P, Aarhus University Hospital, Tage Hansens-Gade, DK-8000 Aarhus, Denmark; lene.h.iversen{at}


Background: The lymph node harvest from colorectal specimens is pivotal for patients with colorectal cancer (CRC), independent of N stage.

Aims: To determine whether the use of GEWF solution (glacial acetic acid, ethanol, distilled water and formaldehyde) could improve the lymph node harvest in CRC specimens.

Methods: Consecutive fresh colonic (n = 60) and rectal (n = 60) specimens from patients with primary CRC resected at Aarhus University Hospital THG between March 2006 and July 2007 were randomised to either conventional preparation or GEWF preparation and examined in a standard manner.

Results: For colonic as well as rectal specimens, the GEWF solution increased the mean lymph node harvest from 9 and 10 to 16 and 17 lymph nodes per specimen compared to conventional prepared specimens (p<0.001). Using the recommended threshold of 12 lymph nodes to ensure adequacy of nodal harvest, the adequacy increased from less than half to almost three quarters independent of tumour origin (p<0.037). The proportion of node-negative specimens was not significantly different between the two preparation groups.

Conclusion: The use of GEWF solution in patients with CRC significantly increases the lymph node harvest of resected specimens.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Accurate assessment of lymph node (LN) status in the resected specimen is essential in the staging of colorectal cancer (CRC) and is pivotal for the decision as to whether postoperative adjuvant therapy should be recommended. Adequate nodal staging is dependant on LN harvest,1 which shows great variability.2 Biological factors such as the individual patient3 and type of specimen4 5 influence this variability, as well as the skill and enthusiasm of the surgeon and the pathologist,3 4 whether the surgeon and the pathologist belong to a multidisciplinary team (MDT) or not,3 the extent of surgical resection, and the extent of pathological examination. The latter depends on the method of assessment.

The conventional method of LN assessment includes serial sectioning, inspection, palpation of specimen being fixed in 10% formalin solution, and microscopic examination after paraffin embedding. Minor LNs in particular, however, may be missed by this method.

To facilitate LN identification in formalin-fixed fatty tissue, various LN revealing solutions for fat clearance have been developed, including alcohol and xylen,6 Carnoy’s solution with chloroform, Koren solution7 and acetone.810 Most techniques are time-consuming, require special equipment and may use noxious chemicals. Such methods are not only costly, but may also defer accurate and timely reporting of stage and nodal status.

In 2001, Newell et al published a method using GEWF solution (glacial acetic acid, ethanol, distilled water and formaldehyde).11 For this method, the specimens were fixed in 10% formalin solution for 16–24 hours and afterwards the pericolic fat was immersed in GEWF solution for 12–18 hours. Following this preparation, the LNs were identified as firm white nodules. In a retrospective study of 67 CRC resections, the mean number of LN per case increased to 10.2 (SD 4.9) compared to 6.8 (SD 3.9) using the conventional method.11 The method is inexpensive and simple to perform.

We conducted a randomised study to evaluate the LN harvest using GEWF solution compared to the conventional methods in CRC resections.


Eligibility criteria

The study comprised specimens from consecutive patients with a primary CRC subjected to a resection at the Department of Surgery P, Aarhus University Hospital, from March 2006 to July 2007. Our department is a referral centre for patients with colonic and non-advanced rectal cancer (admission area: ∼350 000 inhabitants) and advanced rectal cancer (admission area: ∼500 000 inhabitants). Thus, the proportion of patients with advanced rectal cancer is relative high. These patients receive preoperative long course radiotherapy (45–60 Gy) according to the Danish Guidelines.12 Surgeons and pathologists attend weekly MDT meetings on rectal cancer patients.


The primary end-point was the LN harvest in each specimen. Secondary end-points were number of metastatic LNs detected and adequate LN harvest by the two methods. Finally, we examined whether the size of the mesentery could be correlated to the number of LNs.

Power calculation

The mean number of LNs in rectal cancer specimens analysed at our pathological department has been shown to be 8.4 in 47 specimens from 2001–02 (Jens Laurberg, personal communication). The threshold for an adequate LN harvest was set at 12 to ensure correct LN staging,13 thus we wanted to detect a difference of at least 4 LNs between the preparation methods. To have a 90% chance of detecting significance at the two sided 5% level and using the standard deviation reported by Newell et al,11 30 specimens in each preparation group were required. Due to differences in size of colonic and rectal mesenteries, colonic and rectal specimens should be analysed separately, thus 60 colonic and 60 rectal specimens were required.


The colonic and rectal specimens were randomly assigned to either conventional preparation with neutral buffered formalin 10% or to GEWF preparation (pre-prepared solution from Bie og Berntsen, Denmark (nr. LAB40914.5000): ethanol 56%, formaldehyde 4%, acetic acid 7%, less than 2% methanol). We randomised the colonic cancer specimens in blocks of four by type of surgery: (advanced) right hemicolectomy, transverse colonic resection, left hemicolectomy, and sigmoid resection to ensure a close balance of the preparation methods among specimens of different origins. Rectal cancer specimens were randomised in blocks of four by preoperative radiotherapy: no preoperative radiotherapy, 5×5 Gy, and long course radiotherapy. Because preoperative long course radiotherapy may reduce the LN harvest5 we decided to enrol only three blocks of four specimens among this specific patient group. The randomisation code was written on paper pasted together, thus concealing the randomisation code, and hidden in numbered sealed envelopes including four such coding papers. All coding papers from the individual envelopes were allocated sequentially before another envelope was opened. One single dedicated nurse enrolled and allocated all specimens. The allocation was done immediately after the dedicated nurse received the specimen from the surgeon. The pathologists were aware of the allocation.

Pathological examination

All colonic specimens were opened immediately, rinsed, and fixed in the allocated solution. Rectal specimens were rinsed and filled with and fixed in the allocated solution.

Conventional and GEWF preparation: The specimens were fixed immediately in 10% formalin solution or GEWF solution, respectively, for at least 16 hours and usually 48 hours.

Identification of LNs was identical for the two preparations and was performed following serial sectioning, inspection and palpation of specimen. Each identified LN was divided if necessary, and each node was embedded in separate blocks, or if too large, divided in more blocks, noting this in the macroscopic evaluation. From each paraffin block one H&E-slice was prepared for microscopic evaluation.

We introduced the GEWF solution preparation at our pathological department in October 2002 and subsequently applied it to all CRC resections. The learning curve of the pathologists concerning the GEWF method should then have been stabilised.

Estimating size of the mesentery

The nurse, who allocated all the specimens, measured the size of the fresh specimens (before any preparation) by measuring the length of the bowel (antimesenteric border), the length of the mesentery at the site adjacent to the bowel, the height of the mesentery (from the antimesenteric site of the bowel to the ligation of the tumour draining vessel), and the areal using point calculation paper (225 points per 100 cm2). If the nurse could not estimate the size of the mesentery, the specimen was excluded.


Since this study was a quality assurance study, ethics committee approval was not required. Approval was obtained from the Danish Data Protection Agency.

Statistical analysis

Results are expressed as means with 95% CIs where appropriate. The mean number of LNs was compared by use of an independent-samples t-test. Non-parametric data were compared using the Mann–Whitney test and Kruskal–Wallis test. Any association between number of LNs and size of the mesentery was tested by computing Spearman’s rho. All analyses were performed separately for colonic and rectal specimens. A p-value <0.05 was considered significant.

Statistical analysis was performed with the use of SPSS V.8.0.2.


Specimen flow

In total, 214 specimens were assessed for eligibility, of which 60 colonic and 60 rectal specimens were allocated for either conventional preparation solution or GEWF solution (fig 1).

Demographic data

The blocking randomisation was successful as type of surgery was equally distributed among the two preparation groups for colon cancer, as was preoperative radiotherapy for rectal cancer. Type of surgery was also equally distributed among the two groups for rectal cancer as well as distance to the anal resection line and to the circumferential resection margin (table 1).

Table 1 Demographic data and tumour characteristics of colonic and rectal specimens prepared in conventional solution and GEWF solution

The conventional and GEWF preparation groups were comparable regarding patients’ age and sex distribution, tumour characteristics, and status of the surgeon and the pathologist (a single consultant examined 92 of the specimens equally distributed among the two groups). None of the measures of the fresh mesentery were influenced by the preparation method (table 2).

Table 2 Size of mesentery of colonic and rectal specimens prepared in conventional solution and GEWF solution

Lymph node harvest

For colonic as well as rectal specimens, the GEWF solution lead to a significantly higher LN harvest per specimen compared to the conventional solution (table 3). The mean LN harvest increased by 7 LNs per specimen independent of tumour origin. Even when excluding the 12 rectal cancer specimens which had received preoperative long course radiotherapy, the LN harvest in rectal specimens prepared with GEWF was significantly higher: mean 17 (95% CI 14 to 21) versus 10 (7 to 13), p = 0.021. The number of metastatic LNs did not differ significantly, as well as the proportions of node-negative and node-positive specimens, between the two preparation groups.

Table 3 Lymph nodes harvest in colonic and rectal specimens prepared in conventional solution and GEWF solution

An adequate LN retrieval was achieved in less than half of the specimens prepared by the conventional method compared to almost three quarters among GEWF prepared specimens (p<0.037), using a threshold of 12 LNs to ensure an accurate N-staging.

Size of mesentery correlated to number of lymph nodes

None of the measures of the mesentery size were significantly correlated to the LN harvest (data not shown).


Adequate assessment of LN status in resected CRC specimens is pivotal for (i) the allocation to postoperative adjuvant chemotherapy and (ii) overall survival.1417 The pathologist’s diligence in pathological examination is one of several factors influencing the LN harvest. The LN assessment can be facilitated by use of LN revealing solutions. In this randomised study, we have shown that by using the GEWF solution, the LN harvest increased significantly in colonic cancer as well as rectal cancer specimens compared to specimens prepared with conventional formalin solution.

Newell et al reported the significance of GEWF solution in 2001 based on a retrospective study of 67 colonic cancer specimens.11 A study using another LN revealing solution including 5% glacial acetic acid also showed significantly higher LN harvest compared to a historical control group.18 Otherwise, the significance of GEWF has not been reported and, to our knowledge, the effect of LN revealing solutions has not previously been evaluated in randomised trials.

It has been stated that 12 negative LNs is appropriate to confer node negative status. This measure was achieved among less than half of conventionally prepared specimens compared to almost three quarters among GEWF prepared specimens. Thus, the GEWF decreases the risk of understaging the nodal stage significantly and may have important implications in the management of the patients. Understaging can lead to metastatic LNs being missed and excluding the patients from beneficial adjuvant chemotherapy, in addition to decreased overall survival.14 15 17

Although the GEWF method increased the LN harvest significantly by 7 LNs per specimen to 16–17 LNs per specimen, the identification of additional LNs had no significant effect on the proportion of node-positive specimens in the present study. Even the numbers of metastatic LNs were comparable. A study of 7062 specimens showed that the percentage of node-positive specimens increases with the number of nodes retrieved.3 However, our study is underpowered to test this specific outcome.

Length of specimen has been reported to correlate with LN harvest in a large scale study.4 We were unable to identify any statistically significant association between length of bowel, size of mesentery and LN harvest in this small scale study.

The strength of this study is the randomised design, resulting in equal distribution of other factors influencing LN harvest among the preparation groups, and it is well powered. The main limitation of our study is that the pathologists could not be blinded for the assignment as GEWF has its own characteristic macroscopic appearance, which is impossible to hide. The single pathologist who examined 92 of the 120 specimens states, however, that her time spent was unaffected by the preparation method. Therefore, we do not believe that a biased LN detecting method favouring the GEWF specimens can explain our findings. In the study of Newell et al,11 the specimens were fixed in formalin 16–24 hours before addition of GEWF. Our set up was changed, because the specimens were immediately fixed in the allocated solution formalin or GEWF. By this means we managed to have complete control of allocation.

In conclusion, using the GEWF preparation of colonic and rectal specimens significantly increased the LN harvest from 9 and 10 to 16 and 17 LNs per specimen compared to conventional prepared specimens.

Take-home messages

  • The use of GEWF solution (glacial acetic acid, ethanol, distilled water and formaldehyde), compared with conventional formalin solution, increased the lymph node harvest significantly in colonic cancer as well as rectal cancer specimens.

  • The GEWF solution may have diagnostic and therapeutic consequences for colorectal cancer patients, since adequate assessment of lymph node status in resected colorectal cancer specimens is pivotal for the allocation to postoperative adjuvant chemotherapy and overall survival.


The authors thank Birgitte Gustafson for her meticulous work. She evaluated all the patients for eligibility to the study, collected and randomised and measured all the specimens, and collected all data in a database.



  • Competing interests: None.