Washout after lobectomy: is water more effective than normal saline in preventing local recurrence?

Abstract

A best evidence topic in thoracic surgery was written according to a structured protocol. The question addressed was: ‘is water washout more effective than normal saline washout after lobectomy in preventing local recurrence?’ Altogether more than 48 papers were found using the reported search, of which nine represented the best evidence to answer the clinical question. The authors, journal, date, country of publication, patient group studied, study type, relevant outcomes and results of these papers are tabulated. Tumour cell ‘spillage’ after cancer resection is linked to a worse prognosis, so washout to minimize contamination is an established surgical technique. While the mechanical effects of lavage are well validated, the differential cytocidal effects of water versus saline as irrigation fluids are not. There are currently no studies addressing this issue in the thoracic surgery setting, after lung cancer lobectomy. However, the majority of relevant papers describe the use of basic in vitro methods and animal models to produce data that can conceivably be extrapolated to the clinical question in hand. The number of studies is small, and some have technical limitations. While two of the better-designed experiments suggest that water exerts a superior cytocidal effect on tumour cells, data from other studies are somewhat unimpressive, with two studies reporting that water washout controls tumour growth to a lesser extent than saline. This, together with the complete paucity of clinical trials on the subject, leads us to conclude that water is unlikely to represent a superior irrigation fluid in lung cancer patients after lobectomy.

INTRODUCTION

A best evidence topic was constructed according to a structured protocol. This is fully described in the ICVTS [1].

THREE-PART QUESTION

In [lung cancer patients undergoing lobectomy], is [water superior to normal saline] in [preventing recurrence]?

CLINICAL SCENARIO

When assisting in theatre, you notice that after lobectomy in lung cancer patients your consultant performs the washout of the chest cavity with water, not normal saline. Apparently, this is a time-honoured tradition because water is more likely to lyse free tumour cells due to its hypotonicity. Although the theory sounds right, you are unsure whether there is evidence to support this practice and resolve to check the literature.

SEARCH STRATEGY

Medline 1948 to September week 4 2011 using OVID interface.

(Lobectomy.mp OR Lung resection.mp OR exp Pneumonectomy/ OR exp Surgery/) AND (exp therapeutic irrigation/ OR Irrigation.mp OR washout.mp) AND (cancer.mp. or exp Neoplasms/ OR Tumour.mp).

SEARCH OUTCOME

Forty-eight papers were found using the reported search. From these, nine papers were identified that provided the best evidence to answer the question. These are presented in Table 1.

Table 1:

Best evidence papers

Author, date, country, study type	Study group	Key outcomes	Key results	Comments
Kosuga et al., 2011, Japan [2], Uncontrolled in vitro and in vivo (human patients) study	Three human oesophageal SCC (ESCC) cell lines: exposed to distilled water	Cell morphology changes	Cell swelling then rupture within 3 min	The most recent study and the only one to focus on pleural lavage. No controls were used, and most values are displayed graphically rather than stated. Although the clinical component appears to be an afterthought, it refutes Huguet et al.'s notion that intrapleural/peritoneal secretions raise the osmolality of lavage fluid such that tumour cell lysis is unlikely to occur. This may encourage the use of pleural lavage in clinical practice, pending higher-quality evidence
		Cell volume changes	Overall cell volume post-hypotonic shock was smaller than pre-hypotonic shock volume, implying cell fragmentation
		Number of ESCC cells after exposure to distilled water	No surviving cells 48 h after 10-min exposure to water in two of the cell lines; the third was more resistant to cytocidal effects and 30-min exposure was needed to kill all cells
	Five patients undergoing surgery for ESCC: received pleural lavage with distilled water	Osmolarity of pleural lavage fluid during surgery	7.6 ± 1.0 (mean ± SEM) mOsmol/kgH2O
Ito et al., 2011, USA [3], Controlled in vitro and in vivo (animal peritoneal lavage model) study	Murine ‘colorectal’ cancer cell line: exposed to 2 ml of water or normal saline for varying periods	Cell morphology changes	Water: immediate changes leading to lysis of >95% cells within 15 min Saline: no cell membrane, cytosol or nuclear changes noted	A well-designed study with adequate controls for both the in vitro and in vivo experiments. The inclusion of data on comorbidities and survival further adds to this paper's relevance to clinical practice
	BALB/c mice: injected intraperitoneally with colorectal cancer cells. Laparotomy incision made and 3 ml of water or normal saline distributed intraperitoneally for 15 min with agitation. Control mice underwent ‘sham’ laparotomy without lavage	Tumour burden on MR imaging at day 23	Water-treated mice: 316 ± 181 mm3 Saline-treated mice: 1477 ± 181 mm3 (P < 0.05) Control mice: 924 ± 162 mm3 (P < 0.05)
		Comorbidities/survival	Water-treated mice: mean survival of 41 days before morbidities were noted Saline-treated mice: all required euthanasia by day 30 due to comorbidities
Hah et al., 2011, South Korea [4], Controlled in vitro and in vivo (animal wound model) study	Syngeneic ‘SCC’ cell line: exposed to distilled water and povidone-iodine (different concentrations and exposure times)	In vitro tumour growth: analysis of survival fraction after 72 h	100% water: tumour growth almost completely inhibited after 5 min exposure 100% povidone-iodine: tumour growth completely inhibited after 30 s exposure	The cell line studies in this investigation did not include normal saline as a control. Fortunately, the murine wound model compared water, saline and povidone-iodine—but despite stating that povidone-iodine and water exert cytocidal effects compared with saline, the authors fail to present specific values
	Twenty-five C3H/HeJ mice: surgical wounds created and incubated with cancer cells for 30 min. Wounds irrigated for 5 min with water or normal saline (or povidone-iodine, hydrogen peroxide or cisplatin). Wounds then sutured closed	In vivo tumour growth	Water had a limited inhibitory effect on tumour growth compared with saline, which had no effect. Povidone-iodine significantly inhibited tumour growth
Lin et al., 2006, China [5], Controlled follow-up study (level III)	32 patients with spontaneous ruptured ‘hepatocellular carcinoma’ undergoing curative liver resection: 13 patients received post-operative peritoneal lavage with distilled water, 19 did not	Number of patients with tumour recurrence	Lavage: 2 patients No lavage: 11 patients	This clinical study effectively compared the benefits of lavage with no lavage in terms of survival time post-liver resection for ruptured hepatocellular carcinoma. The differential effects of water versus normal saline as irrigation fluid are not investigated
		Mean disease-free time	Lavage: 3.59 ± 0.60 years No lavage: 2.05 ± 0.74 years (P = 0.045)
		Survival time	P = 0.0158
Huguet et al., 2004, UK [6], Controlled in vitro study and uncontrolled in vivo (human patients) assay	Five patients immediately post- ‘colorectal cancer’ resection surgery: received five sequential 1000-ml peritoneal lavages. Each aliquot left to equilibrate for 1 min with agitation. Sample of fluid taken before aspiration and administration of next aliquot Two animal colorectal cancer cell lines: exposed to water or saline (or betadine or culture medium)	Osmolality of lavage water	50 mM after first lavage; decreased with each subsequent lavage, reaching plateau of 10 mM after third lavage	This oft-cited study sought to comment on the utility of peritoneal lavage as it was being performed by surgeons at the time. The authors found that three sequential lavages were required to dilute lavage fluid to an osmolality of 10 mM, since it becomes contaminated with intra-peritoneal detritus upon introduction into the cavity. More significantly, the authors found that at this minimum osmolality, more than 30 min are required for lavage fluid to achieve 100% tumour cell lysis in vitro. This suggests that current lavage techniques are not in fact efficacious
		Cell lysis (number of cells surviving intact)	Water: 100% cell lysis within 12 min Saline: 0% cell lysis within 36 min Linear regression analysis—  water more toxic than saline (P < 0.0001) Betadine: 100% cell lysis not achieved within 40 min Linear regression analysis—water more toxic than betadine (P < 0.0001) ‘Lavage solution’ (10 mM): 100% cell lysis within 32 min Linear regression analysis—  ‘lavage solution’ more toxic than saline (P < 0.0001) and betadine (P < 0.0008) but less toxic than water (P < 0.0001)
Allegretto et al., 2001, USA [7], Controlled in vivo (animal wound model) study	40 non-obese diabetic/SCID mice: surgical wounds created and inoculated with human ‘SCC’ cells. Ten mice had incisions closed immediately (controls), 10 had water irrigation, 10 had saline irrigation and 10 had 1 mM gemcitabine irrigation	Percentage of mice with palpable tumour growth	Day 17: ‘control’ −70%, ‘water’ −15% (P < 0.0005), ‘saline’ −0% (P < 0.0001), ‘gemcitabine’ −0% Day 56: ‘control’ −80%, ‘water’ −75%, ‘saline’ −40% (P < 0.01), ‘gemcitabine’ −35% (P < 0.04)	This wound model study is inferior to that by Ito et al., as the outcome measure is relatively subjective. However, adequate controls were used and the authors presented data and significance values in a transparent manner. Interestingly, results indicate that saline is superior to water for long-term control of tumour growth
Morris and Scholten 1996, USA [8], Controlled in vivo (animal model) study	36 nude mice: injected intraperitoneally with human ‘ovarian cancer’ cells. Intraperitoneal lavage administered with saline or water. Mice were followed up until they reached moribund status or until 60 days post-injection	Percentage of mice with clinically detectable tumour growth at 30 days	Water: 55%, saline: 89% (P < 0.03)	This peritoneal lavage model study again uses ‘soft’ outcome measures to compare water with saline lavage in an ovarian carcinoma setting. Data for tumour recurrence at 30 days are in favour of water over saline, but the other longer-term outcomes suggest that saline is in fact significantly superior to water in controlling the sequelae of tumour spillage
		Ascites	Water: 72%, saline: 22% (P = 0.03)
		Composite tumour score at 60 days (number and size of tumour nodules at dissection)	Water: 4, saline: 2 (P = 0.05)
		Time to moribund status requiring euthanasia	‘Water lavage’ mice required euthanasia earlier than ‘saline lavage’ mice (P = 0.022)
Sweitzer et al., 1993, USA [9], Controlled in vivo (animal wound model) study	C57BL/6 mice: surgical wounds created and inoculated with syngeneic B16-F10 ‘melanoma’ cells. Mice received saline irrigation, water irrigation or no irrigation	Tumour weight at dissection (g)	Water: 0.0 ± 0.0 Saline: 0.007 ± 0.015 No irrigation: 0.017 ± 0.012 (P < 0.03)	This well-controlled study investigated the effects of saline versus water versus no irrigation. The authors also incorporated experiments to test the effects of: 1. Timelag between tumour inoculation and lavage 2. Duration of incubation with lavage fluid 3. Concentration of tumour cells per unit volume of inoculating fluid There were no significant differences reported for any of the above factors, with the exception that irrigation with either fluid was beneficial in the ‘immediate irrigation’ setting after contamination with a low-concentration inoculum
Park et al., 1991, UK [10], Controlled in vitro study	Two human ‘breast cancer’ cell lines: exposed to water or phosphate-buffered saline (or Savlon, noxythiolin, povidone-iodine, hydrogen peroxide or bleomycin)	Number of viable cells	Water: ∼50 PBS: ∼130 (P < 0.01)	This early study compared six irrigation fluids to a control fluid, PBS. Normal saline was not used. Nevertheless, the simple outcome measure suggests that water has a significantly greater tumoricidal effect than PBS

PBS, phosphate-buffered saline.

RESULTS

Park et al.'s in vitro study compared the cytotoxic effects of different irrigation fluids. The authors validated tryptan blue as an accurate marker of breast cancer cell viability, then showed that water reduced the number of viable tumour cells compared with phosphate-buffered saline (P < 0.01). Normal saline was omitted from investigation.

At the other end of the spectrum, Lin et al.'s clinical trial measured recurrence after liver resection for ruptured hepatocellular carcinoma. Peritoneal lavage with water (recurrence in 15%) was superior to no peritoneal lavage (recurrence in 58%). Normal saline again did not feature, rendering it impossible to separate factors related to lavage per se from factors caused by osmotic absorption.

Sweitzer et al. ‘seeded’ a murine wound model with melanoma cells and looked at effects of water versus saline irrigation, timelag to irrigation, duration of irrigation and tumour cell concentration of the inoculum. Overall, there were no outcome differences between mice receiving water, saline, or no irrigation (P > 0.73), suggesting that neither the mechanical action of lavage nor the hypotonic effect of water alter tumour growth.

Allegretto et al. used a similar murine model to assess three irrigation fluids (water, saline, gemcitabine) in the head and neck cancer setting. Mice with non-irrigated wounds served as controls. All fluids initially delayed the development of squamous cell carcinomas (SCCs)—by day 24, 70% of controls had developed tumours versus 15% of the ‘water’ group (P < 0.0005), 0% of the ‘saline’ group (P < 0.0001) and 0% of the ‘gemcitabine’ group (P < 0.0001). By day 56, incidence of recurrence in the ‘water’ group was similar to that of controls. In contrast, saline (P < 0.01) and gemcitabine (P < 0.004) improved tumour control compared with water irrigation or no irrigation.

This result was corroborated by Morris and Scholten using their peritoneal lavage murine model of ovarian tumour control. Although short-term data suggested that water controls recurrence better than saline (55 versus 89%, P = 0.03), longer-term outcomes suggested that water exerts a smaller effect than saline (ascites: P = 0.003, tumour score at dissection: P = 0.05, survival: P = 0.002). However, most endpoints were subject to variable interpretation.

Some papers attempt to correlate the differential effects of water and saline between in vitro and in vivo settings. Like Allegretto's group, Hah et al. used an animal wound model to investigate tumoricidal effects on SCC cells. Preliminary work with cell lines confirmed water's lytic activity, but saline was not used as a control. In vivo experiments showed that water had a small effect on tumour growth, while saline had none.

In Huguet et al.'s appealing study, the authors navigate seamlessly between clinical and experimental settings. They first showed that peritoneal secretions contaminate lavage water after colorectal cancer resection, producing a minimum osmolality of 10 mM. With this in mind, they incubated colorectal cancer cells with water, saline, Betadine and simulated ‘lavage solution’ (osmolality 10 mM). Water was most cytocidal, followed by ‘lavage solution’ and then Betadine. Saline exerted no significant lytic effect compared with water (P < 0.0001). The authors concluded that the efficacy of ‘oncological’ lavage is likely to be reduced by intraperitoneal secretions, recommending sequential lavages and an incubation period of 32 min to maximize cell lysis. They conceded the impracticality of this in the operating theatre.

Ito et al.'s elegant investigation correlated in vitro and animal model data to show that the lytic properties of water translate into superiority over saline in improving the sequelae of colorectal cancer cell spillage at laparotomy. A highlight of this study is its use of magnetic resonance imaging to provide an objective and quantitative outcome measure (tumour volume in ‘water’ mice 316 ± 181 mm³ versus ‘saline’ mice 1477 ± 181 mm³, P < 0.05). The authors also reported significant differences in peritoneal tumour burden, comorbidities and mouse survival.

Kosuga et al.'s recent work stands alone in its focus on cytocidal effects of hypotonic shock in pleural rather than peritoneal lavage, albeit using oesophageal SCC cells. In the in vitro experiments, the authors employed video microscopy and flow cytometry to demonstrate morphological cell changes and loss of cell viability. However, the paper presents mainly graphical data, suffering from lack of clarity on actual values. Furthermore, while solutions of graded hypotonic osmolalities are used to assay cytocidal effects, isotonic fluid such as normal saline is not used as a control. Finally, the authors performed intra-operative pleural lavage on several oesophageal cancer patients, attempting to debunk the notion that its clinical utility may be undermined by contamination with intrapleural secretions and cells. They reported only a small increase in lavage fluid osmolarity of <10 mOsmol/kgH₂O and then showed that cancer cell rupture alone can elevate osmolarity in vitro by 10 ± 6.5 mOsmol/kgH₂O. They concluded that osmolarity of lavage fluid prior to washout has a more significant influence on cytocidal effects than that after washout.

CLINICAL BOTTOM LINE

Tumour cell ‘spillage’ after cancer resection is linked to worse prognosis, so washout to minimize contamination is an established technique. While mechanical effects of lavage are well-validated, differential cytocidal effects of water versus saline are not. Currently, no studies address this issue in the thoracic surgery setting. However, most papers use in vitro methods and animal models to produce data that can conceivably be extrapolated to the clinical question in hand. There are few studies, some with technical limitations. While two of the better-designed experiments suggest that water exerts a superior cytocidal effect on tumour cells, data from other studies are unimpressive, with two studies reporting that water washout controls tumour growth less effectively than saline. This, together with the complete paucity of clinical trials on the subject, leads us to conclude that water is unlikely to represent a superior irrigation fluid in lung cancer patients after lobectomy.

Conflict of interest: none declared.