Secondary Logo

Journal Logo


Endothelial cell damage after cataract surgery: Divide-and-conquer versus phaco-chop technique

Storr-Paulsen, Allan MD; Norregaard, Jens Christian MD, PhD; Ahmed, Shakil MD; Storr-Paulsen, Thomas MD; Pedersen, Tina Hyldebrandt MD

Author Information
Journal of Cataract & Refractive Surgery: June 2008 - Volume 34 - Issue 6 - p 996-1000
doi: 10.1016/j.jcrs.2008.02.013
  • Free


In the early era of phacoemulsification, most surgeons performed the procedure with the nucleus luxated into the anterior chamber, leading to a 20% to 30% loss of endothelial cells during surgery. After the development of capsulorhexis and hydrodissection, phacoemulsification has been performed in the capsular bag or in the iris plane, causing less endothelial cell loss. In the early 1990s, the divide-and-conquer nuclear cracking technique was introduced.1,2 The advantage was safer surgery with less endothelial cell loss.3 In 1993, the phaco-chop technique for cracking the nucleus was described by Nagahara (K. Nagahara, MD, “Phaco-Chop Technique Eliminates Central Sculpting and Allows Faster, Safer Phaco,” Ocular Surgery News, International Edition, October 1993, pages 12–13). Koch and Katzen4 improved the technique, making it safer in eyes with harder nuclei (stop-and-chop phacoemulsification). However, conflicting reports of endothelial cell damage after phaco-chop and divide-and-conquer phacoemulsification have been published.5,6 The present study was undertaken because of this controversy.


This clinical prospective study included 60 eyes of 60 consecutive patients scheduled to have cataract surgery at the eye clinic of Frederiksberg University Hospital, Frederiksberg, Denmark. The study protocol was approved by the local ethics committee and was reported to the Danish Data Protection Agency.

Exclusion criteria were corneal pathology, pseudoexfoliation, history of ocular trauma or intraocular surgery, intraocular inflammation, diabetes mellitus, age younger than 40 years, preoperative pupil dilation less than 4.0 mm, preoperative endothelial cell count less than 1500 cells/mm2, preoperative anterior chamber depth (ACD) less than 2.5 mm, and surgical complications. No patient was taking preoperative eye medication.

The patients were randomly assigned to 1 of 2 groups. Thirty patients had phacoemulsification using the divide-and-conquer technique, and 30 patients had phacoemulsification using the phaco-chop technique. Sample size was based on a power calculation (power 0.90, P = .05) using standard deviations obtained in previous studies from Frederiksberg University Hospital. For the power calculation, a clinically important change in cell count was defined as a loss of 300 cells/mm2.

Preoperatively, all patients had slitlamp and retinal examinations and best corrected visual acuity (BCVA) and intraocular pressure (IOP) measurements. The firmness of the nucleus was graded using the Emery classification.7

Endothelial cell density (cells/mm2), variation in the size of the endothelial cells (CV), the percentage of hexagonal cells, and central corneal thickness (CCT) were analyzed using a noncontact specular microscope (SP 2000P, Topcon) with the Image-Net imaging system (version 2.1, Topcon). Corneal endothelial photographs were taken preoperatively and 3 and 12 months postoperatively. During each visit, 3 photographs of each cornea were taken and analyzed independently by 3 ophthalmologists unaware of the surgical technique used. The mean of the 9 readings was calculated and used as the final reading for the visit.

Surgical Technique

All patients were operated on by the same surgeon (A.S.-P.). The surgical technique was similar in all cases except for the method of nucleus fracturing. The divide-and-conquer technique was according to Gimbel,2 and the phaco-chop technique was according to the guidelines of Nagahara with the use of a Haefliger chopper (K. Nagahara, MD, “Phaco-Chop Technique Eliminates Central Sculpting and Allows Faster, Safer Phaco,” Ocular Surgery News, International Edition, October 1993, pages 12–13). The surgeon followed a standardized procedure: topical anesthesia (lidocaine 2% in gel suspension) on the cornea, a 2.75 mm self-sealing temporal clear corneal incision, intracameral injection of 0.5 mL of preservative-free lidocaine 1% followed by injection of a dispersive ophthalmic viscosurgical device (OVD) (hyaluronate 3.0% [Vitrax]), capsulorhexis, nuclear fracturing (divide-and-conquer or phaco-chop technique), cortical cleanup, implantation of a foldable acrylic intraocular lens (IOL) (AcrySof MA60AC, Alcon Laboratories, Inc.), and aspiration of the OVD. At end of the operation, 4 mL of cefuroxime (2.5 mg/mL) was injected into the anterior chamber and behind the IOL,8 followed by 1 drop of topical apraclonidine (Iopidine). No sutures were applied.

The surgeon used a Storz Millennium phacoemulsification machine (Bausch & Lomb, Inc.) with a Kelman-type microtip (30 degree) and standardized machine settings. In the divide-and-conquer group, a sculpting program (vacuum 20 mm Hg; maximum 40% ultrasound [US] power; bottle height 75 cm) and a pulse-mode program (vacuum 300 mm Hg; maximum 40% US power; bottle height 135 cm) were used. In the phaco-chop group, the sculpting program was bypassed and only the pulse-mode program was used. Cortical cleanup and OVD removal were the same in the 2 groups. Throughout surgery, a balanced salt solution was used as the irrigation solution.

During surgery, the total time of the operation, diameter of the pupil (estimated by calliper), total phaco time (seconds), and mean phaco power (%) were measured. The total phaco energy was calculated by multiplying the phaco time (seconds) by the mean phaco power (%). The volume of irrigation fluid was calculated by weighing the irrigation bottle before and after surgery.

Patients were dismissed from the ambulatory center 15 to 30 minutes after surgery. Postoperative treatment included dexamethasone (1 mg/mL) 3 times daily for 2 to 3 weeks.

Statistical Analysis

Differences between groups in demographic and clinical characteristics were evaluated using chi-square tests for categorical variables and t tests for continuous variables. Presurgery versus postsurgery changes within groups were analyzed using paired t tests. Mean visual acuity was calculated based on logMAR transformation. Visual acuities are presented as Snellen decimal fractions.

Associations between cell loss and age, ACD, method of fracturing the nucleus, firmness of the nucleus, axial length (AL), volume of irrigation solution, and total phaco energy were evaluated using univariate regression analysis.


In 2 patients originally assigned to the phaco-chop group, the technique was changed to divide and conquer due to very hard nuclei. These 2 patients were excluded from the study and replaced by 2 others. The remaining surgeries were uneventful.

Of the 60 patients enrolled, all completed the 3-month follow-up. Three patients in the divide-and-conquer group did not attend the planned 12-month examination. One had died, and 2 did not show up although they received 2 reminder telephone calls. Table 1 shows the baseline preoperative patient demographics and clinical data. There were no statistically significant differences between the 2 groups in age, ACD, AL, keratometry readings, nuclear firmness, visual acuity, or IOP.

Table 1
Table 1:
Baseline preoperative clinical data.

Table 2 shows the surgical parameters. There were no differences in the mean size of the pupil, volume of irrigation fluid, or length of surgery. The mean total phaco energy was 3.98 ± 2.5 (SD) in the phaco-chop group and 12.79 ± 8.6 in the divide-and-conquer group. The difference was highly statistically significant (P<.0001).

Table 2
Table 2:
Surgical parameters.

Preoperative cell density was similar in the 2 groups (2742 ± 424 cells/mm2 phaco chop and 2747 ± 330 cells/mm2 divide and conquer) (P = .96). There were no statistically significant differences between the 2 groups in any corneal parameters (Table 3).

Table 3
Table 3:
Preoperative and postoperative corneal endothelial morphology.

Both groups had a significant decrease in endothelial cell density at 3 and 12 months. The mean cell loss was 173 cells/mm2 (6.3%) and 155 cells/mm2 (5.7%) at 3 months and 12 months, respectively, in the phaco-chop group and 138 cells/mm2 (5.0%) and 94 cells/mm2 (3.5%), respectively, in the divide-and-conquer group. The difference between the 2 groups was not significant at either follow-up (Table 3). Table 4 shows the changes in preoperative cell parameters and postoperative values within the groups. There was no significant change in the variance of endothelial cell size (CV), percentage of hexagonal cells, or CCT (P>.05, t test).

Table 4
Table 4:
Changes in cell parameters 3 and 12 months postoperatively.

There was an equal and significant increase in mean Snellen BCVA from preoperatively to 3 months postoperatively, from 0.34 to 0.80 in the phaco-chop group and from 0.40 to 0.90 in the divide-and-conquer group. No change in IOP was noted in either group at any time.

Univariate regression analysis of the associations between cell loss and age, ACD, method of fracturing the nucleus, firmness of the nucleus, AL, volume of irrigation solution, and total phaco energy found that only shorter AL had a positive correlation with higher endothelial cell loss (β = –2.3; standard error = 0.75, P = .0075).


In the present study, we compared endothelial cell damage in cataract surgery between the divide-and-conquer technique and the phaco-chop technique. Preoperatively, demographic data and endothelial cell parameters were very similar between the 2 groups, indicating that no sampling bias was present. Specular microscopy should be performed at least 3 months postoperatively, when cell loss and reorganization have stabilized.9 This guideline was followed in the present study.

We performed the phaco-chop technique as described by Nagahara (K. Nagahara, MD, “Phaco-Chop Technique Eliminates Central Sculpting and Allows Faster, Safer Phaco,” Ocular Surgery News, International Edition, October 1993, pages 12–13). The technique has the disadvantage of providing little room for the nuclear pieces to move during surgery in eyes with a hard nucleus. Koch and Katzen4 modified the phaco-chop technique by starting with central sculpting to provide space and facilitate separation of the nucleus. This modification to Nagahara's original technique was applied to firm nuclei (Emery grade ≥3) in the present study.

All cataract surgery is associated with endothelial cell damage, which is reflected by a decrease in cell density and a change in cell morphology.10 The extent of endothelial damage depends on the incision type, OVD used, type of IOL, composition of the irrigation solution, total phaco energy, and location of active phacoemulsification.11–15

In a prospective study, Pirazzoli et al.6 showed that the phaco-chop technique led to shorter phaco time and less endothelial cell loss than the divide-and-conquer technique. Unfortunately, the follow-up was only 8 weeks. Other studies16,17 report that the phaco-chop technique requires less total phaco energy than the divide-and-conquer technique, although none of the studies evaluated endothelial damage. It has been postulated that less total energy leads to less endothelial cell damage.6,11 This positive correlation was confirmed by some15,18 but not by others.5,13

In most studies, the cell loss after phaco chop was similar to that in the present study, in which the mean loss 3 months postoperatively was 6.3%. However, in studies showing a significant difference in cell loss between the 2 techniques, cell loss as high as 12.0% has been reported after divide-and-conquer phacoemulsification.6 This cell loss is much higher than the 5.0% reported in our study. This decrease in cell loss since 1996 might be due to technical improvements in the phaco tip and OVDs and suggests that the current divide-and-conquer technique is currently as safe as the more recently developed phaco-chop technique.

The results in the present study are in line with those of Zetterström and Laurell5 and Hayashi et al.13 Similar to Zetterström and Laurell, we found the same endothelial cell loss in the 2 groups, although the total phaco energy was significantly less in the phaco-chop group. Hayashi et al. did not find a correlation between phaco energy and endothelial cell loss and suggest that mechanical contact with nuclear fragments could be the principal cause of endothelial injury.

Using a side-view camera video, Davison et al.19 showed that the divide-and-conquer technique delivers more US power behind the iris than the phaco-chop technique. Furthermore, the phaco-chop technique requires a longer period of manipulation of fragments in the anterior chamber than divide-and-conquer surgery, in which most manipulation takes place behind the iris. This may be why we observed similar endothelial cell loss with both methods, despite significantly less total phaco energy. This hypothesis may also be supported by the higher cell loss in patients with a shorter AL and thus a shallower anterior chamber. In a shallow chamber, the phaco tip and the lens fragments are closer to the endothelium.

The present study was not designed to identify risk factors for cell loss as the 2 groups were very similar except for surgical technique. Only shorter AL was positively correlated with endothelial cell loss. This result is similar to that in a study by Walkow et al.15 Others have identified further risk factors. Hayashi et al.13 performed multiple linear regression analysis of possible risk factors for endothelial injury during phacoemulsification in 859 patients. They identified high nucleus grade, greater infusion volume, type of IOL, and larger nucleus as independent predictors of endothelial cell loss in divide-and-conquer phacoemulsification (R2 = 0.42). Walkow et al.15 identified AL and increased phaco time as independent predictors (R2 = 0.39). Using multiple regression analysis, O'Brien et al.18 found that dense cataract (grade 3) and long absolute phaco time were independent predictors of endothelial cell loss (R2 = 0.36). In the present prospective study, the same surgeon performed all procedures using the same phaco machine with standardized settings and the same type of IOL. Furthermore, eyes with cataract of Emery grade 0, 1, or 5 did not have surgery because grade 0 and 1 represent too mild a cataract and grade 5 is seldom seen in Denmark. In the present study, the mean nucleus firmness was similar in the 2 groups (P = .43) and was lower than grade 3.

In conclusion, we found no difference in endothelial cell loss between the 2 surgical techniques, suggesting that phacoemulsification using the current divide-and-conquer technique is as safe as the more recently developed phaco-chop technique.


1. Shepherd JR. In situ fracture. J Cataract Refract Surg. 1990;16:436-440.
2. Gimbel HV. Divide and conquer nucleofractis phacoemulsification: development and variations. J Cataract Refract Surg. 1991;17:281-291.
3. Hayashi K, Nakao F, Hayashi F. Corneal endothelial cell loss after phacoemulsification using nuclear cracking procedures. J Cataract Refract Surg. 1994;20:44-47.
4. Koch PS, Katzen LE. Stop and chop phacoemulsification. J Cataract Refract Surg. 1994;20:566-570.
5. Zetterström C, Laurell C-G. Comparison of endothelial cell loss and phacoemulsification energy during endocapsular phacoemulsification surgery. J Cataract Refract Surg. 1995;21:55-58.
6. Pirazzoli G, D'Eliseo D, Ziosi M, Acciarri R. Effects of phacoemulsification time on the corneal endothelium using phacofracture and phaco chop techniques. J Cataract Refract Surg. 1996;22:967-969.
7. Emery JM., 1983. Kelman phacoemulsification; patient selection. In: Emery JM, McIntyre DJ, editors., Extracapsular Cataract Surgery. CV Mosby, St Louis, MO, pp. 95-100.
8. ESCRS Endophthalmitis Study Group. Prophylaxis of postoperative endophthalmitis following cataract surgery; results of the ESCRS multicenter study and identification of risk factors. J Cataract Refract Surg. 2007;33:978-988.
9. Schultz RO, Glasser DB, Matsuda M, Yee RW, Edelhauser HF. Response of the corneal endothelium to cataract surgery. Arch Ophthalmol. 1986;104:1164-1169.
10. Bourne RRA, Minassian DC, Dart JKG, Rosen P, Kaushal S, Wingate N. Effect of cataract surgery on the corneal endothelium; modern phacoemulsification compared with extracapsular cataract surgery. Ophthalmology. 2004;111:679-685.
11. Dick HB, Kohnen T, Jacobi FK, Jacobi KW. Long-term endothelial cell loss following phacoemulsification through a temporal clear corneal incision. J Cataract Refract Surg. 1996;22:63-71.
12. Holzer MP, Tetz MR, Auffarth GU, Welt R, Völcker H-E. Effect of Healon5 and 4 other viscoelastic substances on intraocular pressure and endothelium after cataract surgery. J Cataract Refract Surg. 2001;27:213-218.
13. Hayashi K, Hayashi H, Nakao F, Hayashi F. Risk factors for corneal endothelial injury during phacoemulsification. J Cataract Refract Surg. 1996;22:1079-1084.
14. Joussen AM, Barth U, Çubuk H, Koch H-R. Effect of irrigation solution and irrigation temperature on the cornea and the pupil during phacoemulsification. J Cataract Refract Surg. 2000;26:392-397.
15. Walkow T, Anders N, Klebe S. Endothelial cell loss after phacoemulsification: relation to preoperative and intraoperative parameters. J Cataract Refract Surg. 2000;26:727-732.
16. DeBry P, Olson RJ, Crandall AS. Comparison of energy required for phaco-chop and divide and conquer phacoemulsification. J Cataract Refract Surg. 1998;24:689-692.
17. Ram J, Wesendahl TA, Auffarth GU, Apple DJ. Evaluation of in situ fracture versus phaco chop techniques. J Cataract Refract Surg. 1998;24:1464-1468.
18. O'Brien PD, Fitzpatrick P, Kilmartin DJ, Beatty S. Risk factors for endothelial cell loss after phacoemulsification surgery by a junior resident. J Cataract Refract Surg. 2004;30:839-843.
19. Davison JA, Cionni RJ, Snowdon JR, Swartz K, Wears S, Simmons S, Barker A, Brother J. Simultaneous surgeon and side-view video analysis comparing in situ fracture and stop-and-chop phacoemulsification. J Cataract Refract Surg. 2005;31:274-279.
© 2008 by Lippincott Williams & Wilkins, Inc.