Symptom Relief and Quality of Life after Combined Partial Hepatectomy and Cyst Fenestration in Highly Symptomatic Polycystic Liver Disease

Abstract

Background:

Polycystic liver disease (PLD) can cause severe symptomatic hepatomegaly. Combined partial hepatectomy and cyst fenestration (PHCF) can be performed to reduce liver volume and symptom burden. We aimed to assess change in symptom relief and quality of life six months after PHCF in PLD patients.

Method:

We established a prospective cohort between 2014 and 2018 at a referral center in the United States. Patients who underwent PHCF for volume-related symptoms were included. Primary outcome was change in PLD-related symptoms, measured with PLD Questionnaire (PLD-Q). Secondary outcomes were change in liver volume (CT/MRI) and change in quality of life, measured with the 12-Item Short Form Survey (SF-12) and the EuroQoL Visual Analogue Scale (EQ-VAS). Questionnaire scores range from 0 to 100 and were assessed before and six months after PHCF. Surgical complications were scored according to Clavien-Dindo (grade 1 to 5).

Results:

We included 18 patients (mean age 52 years, 82% female). PHCF reduced median liver volume (4917 to 2120 ml). Symptoms, measured with PLD-Q, decreased (76.9 to 34.8 points; p<0.001) six months after surgery; 15/16 symptoms declined after treatment, with most impact seen on early satiety and dyspnea. Quality of life also improved after surgery: median physical and mental component scales of the SF-12 and EQ-VAS increased (24.9 to 45.7, p=0.004; 40.5 to 55.4, p=0.02; and 40.0 to 72.5, p=0.003). Major complications (grade 4) occurred in two patients. There was no procedure-related mortality.

Conclusion:

PHCF substantially improves symptom burden and quality of life in highly symptomatic PLD patients.

TOC Statement-

We prospectively investigated symptom relief and quality of life following partial hepatectomy and cyst fenestration (PHCF) in highly symptomatic polycystic liver disease patients. We found PHCF to substantially improve symptom burden and quality of life in this population.

Introduction

Polycystic liver disease (PLD) is a rare hereditary condition that results in progressive hepatomegaly in a proportion of patients.¹ It is associated with autosomal dominant polycystic kidney disease (ADPKD), and occurs less commonly without renal cysts in autosomal dominant polycystic liver disease (ADPLD).² The enlarged cystic liver may cause compression of adjacent abdominal structures and organs, leading to symptoms and a compromised quality of life.³ This underscores the need for volume-reducing therapies that improve patient-reported outcomes.

There are limited therapeutic options for patients with severe polycystic liver disease. Except for liver transplantation, there is no curative treatment available yet. However, PLD patients usually require exception points for transplantation, since liver function remains preserved.⁴ Aspiration sclerotherapy and laparoscopic fenestration are limited to patients with superficial large cysts, and while reduction of symptoms and improvement of quality of life can be achieved, the rate of recurrent symptoms and need for re-intervention is high.^5–7 For diffuse cystic involvement, transcatheter arterial embolization (TAE) is a novel intervention to reduce cyst volume, but more experience by Western groups is needed to confirm the safety and efficacy of this procedure.^{8, 9} Somatostatin analogues curtail liver cyst growth in clinical trials and improve quality of life, but the proportional reduction in liver volume is limited to 5% after one year.^{10, 11}

Currently, large liver volume reduction in PLD can only be achieved by combined partial hepatectomy and cyst fenestration (PHCF).¹² A large retrospective study (n=186) reported a median liver volume reduction of 61% after PHCF that was durable up to 20 years.¹³ This establishes PHCF as an effective volume-reducing procedure. Despite being the primary goal of treatment, the effectiveness of PHCF on symptom relief and quality of life remains to be elucidated.

Results from a single study suggested that quality of life after PHCF is comparable to that of healthy controls after a mean follow-up of 9 years.¹² However, baseline scores were missing, no validated symptom questionnaire was used and follow-up duration varied between patients.¹² Furthermore, it should be investigated whether this procedure leads to a clinically relevant change in symptoms and quality of life that outweighs the risk of surgery.

The aim of this study was to prospectively investigate symptom relief and improvement of quality of life using validated questionnaires in a cohort of PLD-patients six months after PHCF. Our secondary aim was to identify the threshold for minimal clinically important differences in PLD-related symptoms and to explore potential predictors for clinically relevant improvement.

Methods

Study design and setting

We performed a prospective cohort study at the Mayo Clinic, Rochester MN, United States (US) between November 2014 and May 2018. Patients received a questionnaire set before surgery and six months after surgery. Non-responders received a reminder call after 2 weeks. Patients that did not respond on paper were asked to complete a phone interview 4 weeks after the initial follow-up survey was sent. This study was approved by the Mayo Clinic Rochester Institutional Review Board (IRB# 14–003832) and informed consent was obtained from all study participants. This study is reported in compliance with the STROBE guidelines for cohort studies (Supplementary Table 1).¹⁴

Procedure

PHCF consists of (extended) hemihepatectomy or multiple segmentectomies, in combination with fenestration of remnant cysts.¹² It is often combined with cholecystectomy for technical reasons.¹⁵ Sometimes PHCF is combined with renal surgery in ADPKD patients (e.g. nephrectomy) for additional cyst volume reduction. All procedures were performed by open approach by two surgeons (RLS, DMN). Preoperative evaluation and surgical technique have been detailed previously.^{12, 13, 15}

Participants

We included patients who underwent PHCF for massive and symptomatic PLD. Patients were selected for surgery by their treating surgeon independent of this study. PHCF was only offered if one or more hepatic sections (sector) were relatively spared of cysts, with preservation of the majority of parenchyma, if afferent and efferent sectoral hepatic vasculature was patent, and liver function was maintained (Figure 1).^{12, 13}

Figure 1.

Magnetic resonance imaging for polycystic liver disease patients. Panel A: Coronal and axial cross section of a patient that was eligible for partial hepatectomy and cyst fenestration (PHCF). Panel B: Coronal and axial cross section of a patient that was not eligible for PHCF as there were no areas of relatively normal liver parenchyma.

Patients who underwent PHCF for reasons other than control of volume-related symptoms and patients that were predisposed to unreliable questionnaire responses due to mental illness or insufficient knowledge of the English language were excluded.

Outcome measures

Baseline characteristics

We collected baseline demographic variables (patient characteristics, laboratory and treatment variables), and surgery-related parameters (indication for surgery, operating surgeon, perioperative blood loss, duration of surgery, concomitant cholecystectomy, concomitant hernia repair, ICU admission, total duration of admission and rate of any complications) from patient records. All complications were scored according to the Clavien-Dindo classification.¹⁶ Type of surgery was defined using the Brisbane 2000 Nomenclature of Hepatic Anatomy and Resections, however exact classification of the resected segments is not always possible due to the disturbed anatomy in polycystic liver disease.¹⁷

Total liver and kidney volume

All patients underwent magnetic resonance imaging (MRI) or computed tomography (CT) as part of the preoperative evaluation. When follow-up imaging was available, change in liver volume was measured. No standardized imaging measurement method was used as symptoms and quality of life were our primary outcomes, and the large reduction of liver volume after partial hepatectomy surpasses the possible measurement errors.¹⁸ A blinded investigator (M.E.) measured total liver volume (TLV), pre- and post-surgery, using Analyze software (Biomedical Imaging Resource, Mayo Clinic, Rochester, MN).¹⁹ In patients with polycystic kidney disease, we also measured total kidney volume (TKV). Disease severity was calculated by dividing total liver volume by height in meters. PLD was classified as mild (<1600 ml/m), moderate (1600 – 3200 ml/m) or severe (>3200 ml/m) based on previously published cut-off values.²⁰

Symptoms and quality of life

We used a questionnaire set, consisting of the disease-specific Polycystic Liver Disease Questionnaire (PLD-Q)²¹ (primary outcome), and two general quality of life questionnaires: 12-Item Short Form Survey (SF-12)²², and the EuroQoL Visual Analogue Scale (EQ-VAS)²³ (secondary outcomes). .

The PLD-Q was extensively validated in cohort of PLD patients in the USA and the Netherlands.²¹ PLD-Q scores the frequency (6-point Likert scale ranging from 1= never to 6= always) and burden (5-point Likert scale ranging from 0=not at all to 5= a lot) of PLD-related symptoms; abdominal fullness, lack of appetite, early satiety, acid reflux, nausea, pain in rib cage, side, abdomen and back, dyspnea after exertion, limited mobility, tiredness, anxiety about the future, concerns for growing liver, dissatisfaction of the size of abdomen and discomfort with intercourse. It is able to distinguish symptoms from polycystic liver disease and polycystic kidney disease.²¹ A total score can be generated by adding the score of the frequency and burden of all symptoms. This score can be transformed to a score ranging from 0–100 and a higher score indicates a higher symptom burden. The total PLD-Q score was not calculated if >1 question score was missing. The general reference population in the US scored a median of 13 points (IQR: 7 – 22).²¹

Overall quality of life was measured with the frequently used SF-12, a shorter version of the SF-36. In validation studies, the SF-12 provides reproducible results of the summarizing physical component score (PCS) and a mental component scores (MCS) compared to the SF-36.^24–26 Scores range from 0–100 and a higher component score implicates a better physical or mental quality of life. PCS and MCS were only calculated if no question score was missing. Results are compared to norm scores, derived from an age-matched (45 – 54 years) female general reference population from the US, as PLD occurs predominantly in females.¹ This reference population scored a median PCS of 51.61 points and median MCS of 53.48 points.²⁷

The EQ-VAS records patients’ self-rated overall quality of life on a vertical 0–100 scale (‘best imaginable health state’ - ‘worst imaginable health state’).²³ A higher score implicates a better overall quality of life. Results are compared to a norm score, derived from an age-matched (45 – 54) general reference population from the US, as sex-matched scores were not available. The general reference population scored 79.2 points on average.²⁸

Sensitivity analyses of quality of life analyses were performed by excluding patients with major complications (Clavien-Dindo grade 4), renal transplant recipients and concomitant renal surgery.

Clinical relevance

A statistically significant change in symptom score after surgery does not necessarily implicate a clinically relevant improvement for individual patients. Previous literature has shown that half the standard deviation of score changes (change between baseline and follow-up score) is equivalent to the Minimal Clinically Important Difference (MCID).²⁹ Subsequently, the proportion of patients with a change in PLD-Q score (primary outcome) after six months that exceeded the MCID was calculated. Subsequently, baseline characteristics (age, liver volume, change in liver volume, pre-surgical PLD-Q, PCS, MCS, EQ-VAS score, diagnosis, ASA status, previous abdominal surgery, Clavien-Dindo complication grade ≥2) and individual PLD-Q symptom score changes of patients with or without a clinically relevant response were compared in an exploratory analysis to find potential response predicting factors.²⁹

In addition, we calculated the effect size of differences six months after therapy for PLD-Q, SF-12 and EQ-VAS scores. Overall effect size is defined as the change in mean quality of life score, divided by the standard deviation of score changes. Overall effect size was assessed by calculating Cohen’s d for paired-samples, d > 0.2 – 0.5 is defined as a small, d > 0.5 – 0.8 as a medium and d > 0.8 as a large effect size.³⁰

Statistical methods

As literature on change in symptoms and quality of life is lacking, no formal a-priori power analysis was performed. We aimed to include 20 patients because of the rarity of this condition and as large differences after surgery were expected. Descriptive variables are expressed as mean (standard deviation [SD]) for normally distributed data, or median (interquartile range [IQR]) for skewed data. We compared pre- and postoperative values of liver volume, PLD-Q total and individual symptom scores, PCS, MCS and EQ-VAS scores, using the Related-Samples Wilcoxon signed-rank test for non-normally distributed continuous data. A one-sample Wilcoxon signed-rank test was used to determine whether median quality of life scores of patients were different from general reference population norm scores before and after surgery. For responder vs. non-responder subgroup analyses, we used the Independent-Samples Mann-Whitney U test for continuous data or the Chi-squared test for categorical data. P-values are two-tailed and a p-value <0.05 is considered statistically significant. In case of missing data, subjects were only excluded from that particular analysis. All analyses were performed using IBM SPSS Statistics version 25 (SPSS Inc., Chicago, IL).

Results

Of 22 eligible patients, 18 were included in the study (Figure 2). After enrollment, one patient was excluded from analysis because no baseline data were available. Response rate per questionnaire, per timepoint ranged between 76% and 94%. Baseline and follow-up questionnaires were completed in full in 13/17 (76%) patients for PLD-Q, 14/17 (82%) for SF-12 and 12/17 (71%) for EQ-VAS (Supplementary Table 2). Median time to return the questionnaire set was six months after surgery.

Figure 2.

Study flow chart. Small boxes at the bottom show the number of patients with complete data (baseline and follow-up data) for a specific outcome. Abbreviations: PLD-Q: Polycystic Liver Disease Questionnaire. SF-12: 12-Item Short Form Survey. EQ-VAS: EuroQoL Visual Analogue Scale.

Baseline characteristics

Baseline characteristics are presented in Table 1. The included patients had a mean age of 52 years and 82% were female. Included patients had moderate (53%) or severe (47%) PLD, with a median TLV before surgery of 5141.8 ml (IQR: 3773.6 – 7374.2). Included patients with polycystic kidney disease had a median TKV of 452.5 ml (IQR: 337.2 – 1032.3) and no patients were on dialysis. Patients were classified as ASA II (59%) or ASA III (41%) before surgery. Indications for surgery were abdominal discomfort or fullness in 88% of patients; 76% pain, 65% early satiety, 35% fatigue and 24% dyspnea.

Table 1:

Baseline characteristics

Baseline characteristics:		n=17
Age (years), mean ± SD		52.0 ± 9.2
Female sex, n (%)		14 (82%)
Race, n (%)	- White	15 (88%)
	- Asian	1 (6%)
	- Hispanic	1 (6%)
Diagnosis, n (%)	- ADPKD	14 (82%)
	- ADPLD	3 (18%)
ASA status, n (%)	- ASA II	10 (59%)
	- ASA III	7 (41%)
Previous abdominal surgery, n (%)	- All	12 (71%z)
	- Hernia repair	4 (24%)
	- Gynecological surgery	4 (24%)
	- Kidney transplant	3 (18%)
	- Liver cyst fenestration	2 (12%)
	- Bilateral nephrectomy	2 (12%)
Height (cm), median (IQR)		171 (162 – 174)
Weight (kg), median (IQR)		67.6 (60.2 – 73.7)
BMI (kg/m2), median (IQR)		23.7 (22.7 – 25.8)
Total liver volume (ml), median (IQR)		5142 (3774 – 7374)
Total kidney volume (ml), median (IQR)		453 (337 – 1032)
eGFR (ml/min/1,73m2), n (%)	- ≥60	12 (71%)
	- 45 – 59	3 (18%)
	- 30 – 44	1 (6%)
Surgery characteristics:
Estimated blood loss (ml), n (%)	- 1 – 500	11 (65%)
	- 501 – 1000	4 (24%)
	- 1001 – 2000	1 (6%)
	- >2000	1 (6%)
Received packed cells, n (%)		6 (35%)
Number of packed cells*, median (IQR)		2.0 (1.0 – 4.5)
Complications (Clavien-Dindo), n (%)	- None	6 (35%)
	- Grade 1	4 (24%)
	- Grade 2	5 (29%)
	- Grade 3	0 (0%)
	- Grade 4	2 (12%)
	- Grade 5	0 (0%)

Baseline characteristics for 17 included patients.

^*:only patients who received ≥1 packed cells (n=6). ADPLD: autosomal dominant polycystic liver disease. ASA: American Society of Anesthesiologists. BMI: body-mass index. eGFR: estimated glomerular filtration rate.

Surgery characteristics

A detailed surgery description per patient is presented in Supplementary Table 3. Concomitant cholecystectomy was performed in 13/17 (76%) patients and hernia repair in 3/17 (18%). Median duration of surgery was 167 minutes (IQR: 143.5 – 209.0) and mean hospital stay was 8.4 days (SD: 3.3). Perioperative blood loss and complications are presented in Table 1. Post-operative complications occurred in 11/17 patients (65%) and most complications were minor (Grade I or II). However, two patients (12%) had to be admitted to the ICU (grade IV). The first patient with a history of abdominal surgeries had extensive abdominal adhesions and developed complications consisting of biliary leakage which required stent placement in the common bile duct, pneumonia, delirium and bilateral lower extremity deep venous thrombosis. The second patient underwent concomitant renal cyst fenestration and developed a wound infection, acute kidney injury and a right ureteric leak which required ureteric stenting. None of the patients had persistent ascites after surgery. There were no patients with post-operative liver failure.

Change in liver volume

Follow-up imaging was available in 13/17 patients. In some patients, follow up was done by their referring physician. In four of them, no postoperative imaging was performed. Median liver volume was reduced from 4781 ml (IQR: 3303 – 6228) to 2110 ml (IQR: 1589 – 2532) (p=0.001), which translates to a median reduction of 57% (IQR: 45 – 67%). Median baseline liver volume of nonanalyzed patients (n=4) was 7185 ml (IQR: 6211 – 9340). Median time between surgery and follow-up scan was 3 months (IQR: 1 – 7.5).

Quality of life

PLD-Q

After surgery, PLD-symptoms decreased (p<0.001). Median total PLD-Q score fell from 76.9 points (IQR: 41.0 – 83.0) at baseline to 34.8 points (IQR: 17.0 – 43.8) six months after surgery (Figure 3A). The effect size was large (d=1.42). Patients had a significantly higher PLD-Q score compared to the general reference population, both before (76.9 vs. 13.0, p=0.001) and after surgery (34.8 vs. 13.0, p=0.001). An overview of baseline symptom severity derived from the PLD-Q is presented in Supplementary Figure 1.

Figure 3:

Patient-reported outcomes pre-surgery (blue) and six months post-surgery (red). Violin plots show the distribution of scores, median (solid line) and interquartile range (dashed lines) are superimposed. Dotted lines represent the general reference population score. Abbreviations: PLD-Q: Polycystic Liver Disease Questionnaire. SF-12: 12-Item Short Form Survey. PCS: Physical Component Scale. MCS: Mental Component Scale. EQ-VAS: EuroQoL Visual Analogue Scale.

The MCID of the PLD-Q was −11.2 points in this cohort, which results in a clinically relevant difference in 10/13 (77%) patients (Figure 4). Only one patient had an increase in PLD-score (+4.2 points), this ASA III patient, with a history of multiple abdominal surgeries and renal transplant, had severe complications after surgery (Grade IV). Characteristics of patients that did (n=10) or did not (n=3) show clinically relevant response are shown in Table 2. Non-responders had a lower symptom burden and a better health state before surgery compared to responders (PLD-Q: 40.3 vs 79.2, p=0.03; EQ-VAS 55.0 vs 40.0, p=0.04). There were more ASA III patients in the non-responders subgroup (p=0.01).The factors age (p=0.1), baseline liver volume (p=0.6), change in liver volume (p=1.0), baseline scores for PCS and MCS (p=0.2; p=0.4), diagnosis (p=0.4), previous abdominal surgery (p=0.2), and perioperative complications grade ≥2 (p=0.4) were not correlated with clinically relevant response.

Figure 4:

Score change per patient per questionnaire score. Polycystic liver disease questionnaire (PLD-Q) score, Physical Component Scale (PCS) and Mental Component Scale (MCS) of the SF-12 questionnaire and EuroQoL-Visual Analogue Scale (EQ-VAS) are presented. Individual patient results are shown; mean and 95% confidence interval are superimposed. For PLD-Q, the Minimal Clinically Important Difference (MCID) is shown at −16.5 points.

Table 2:

Responder subgroup analysis

Characteristic:	Responders (n=10)	Non-responders (n=3)	P-value
	Median (IQR)	Median (range)a
Age (years)	48 (44 – 57)	60 (52 – 71)	0.1
Liver volume pre-surgery (mL)	5098 (4436 – 7049)	6571 (3961 – 10556)	0.6
Change in liver volume (mL)	2794 (2306 – 5444)	2944 (1427 – 4462)	1.0
PLD-Q pre-surgery (score)	79.2 (64.3 – 84.4)	40.3 (34.0 – 41.7)	0.03
PCS pre-surgery (score)	22.1 (16.8 – 25.8)	25.7 (24.8 – 48.7)	0.2
MCS pre-surgery (score)	40.5 (25.8 – 57.1)	60.2 (22.7 – 67.6)	0.4
EQ-VAS pre-surgery (score)	40.0 (35.5 – 42.5)	55.0 (50.0 – 60.0)	0.04
	N (%)	N (%)
Diagnosis (% ADPKD)	8 (80%)	3 (100%)	0.4
ASA status (ASA II : III)	8:2	0:3	0.01
Previous abdominal surgery (% yes)	6 (60%)	3 (100%)	0.2
Clavien-Dindo complication grade ≥2 (%)	4 (40%)	2 (67%)	0.4

Comparison of responders and non-responders based on a minimal clinically important difference of −11.2 points for the PLD-Q.

^(a.)Total range, as only 3 values were available. Abbreviations: IQR: interquartile range, PLD-Q: polycystic liver disease questionnaire, PCS: physical component scale. MCS: mental component scale. EQ-VAS: EuroQoL-Visual Analogue Scale. ASA: American Society of Anesthesiologists. ADPKD: autosomal dominant polycystic kidney disease.

Nearly all individual symptoms of the PLD-Q declined after treatment (Figure 5). Only change in ‘abdominal pain’ was not statistically significant (p=0.07). This can be explained by an increase of abdominal pain in four patients, which was not associated with major complications or kidney volume change. Early satiety and shortness of breath showed the largest decrease, with a reduction of median score of 6 points. Changes in individual symptom scores for response subgroups are shown in Supplementary Figure 2.

Figure 5:

Radar chart of median individual symptom scores of the Polycystic Liver Disease Questionnaire, pre-surgery (blue) and six months post-surgery (red). *: p<0.05; **: p<0.01

SF-12

Quality of life as assessed by the SF-12 questionnaire improved significantly after surgery. Median PCS increased (p=0.004) from 24.9 points (IQR: 17.4 – 26.5) to 45.7 points (IQR: 34.6 – 55.3) (Figure 3B). Median MCS increased (p=0.02) from 40.5 points (IQR: 25.9 – 59.0) to 55.4 points (IQR: 50.8 – 58.9) (Figure 3C). The effect size was large for increase of PCS (d=0.93) and medium for MCS (d=0.79). Patients had a worse PCS score before surgery compared to the general reference population (24.9 vs. 51.61, p=0.001), while no significant difference remained after surgery (45.7 vs. 51.61, p=0.07). Median MCS was also significantly lower before surgery (40.5 vs. 53.48, p=0.05) and was comparable to the general reference population after surgery (55.4 vs.53.48, p=0.4). Score changes per individual patient are shown in Figure 4. The MCID of the PCS and MCS was +8.4 and +7.6 points in this cohort, respectively. This results in a clinically relevant improvement of physical and mental quality of life in 9/14 (69%) and 8/14 (57%) patients (Figure 4).

EQ-VAS

On a visual analogue scale of quality of life, patients scored a median of 40.0 (IQR: 36.8 – 48.8) before surgery. Six months after surgery, patients scored a median of 72.5 (IQR: 63.3 – 92.3) which was a significant improvement (p=0.003) (Figure 3D, Figure 4). The effect size was large (d=1.89). Patients scored worse before surgery compared to the general reference population (40.0 vs. 79.2, p=0.002) and had a comparable score after surgery (72.5 vs. 79.2, p=0.6). The MCID of the EQ-VAS was +6.9 points, therefore 12/12 (100%) patients had a relevant improvement of health state.

Sensitivity analyses

When excluding patients that suffered major complications (n=2), there were no clinically relevant differences in PLD-Q, PCS, MCS and EQ-VAS score changes (p=0.002, p=0.01, p= 0.04, p=0.003, respectively). Also, results did not change significantly after excluding renal transplant recipients (n=3; p=0.005, p=0.03, p=0.05, and p=0.008, respectively) and patients that had concomitant renal surgery (n=2; p=0.003, p=0.01, p=0.02, and p=0.003, respectively).

Discussion

PHCF significantly improves both symptom burden and quality of life in moderate and severe PLD patients with relative sectorial sparing. Based on our primary outcome (PLD-Q), the majority of patients (77%) met the threshold of a clinically relevant response six months after treatment.

A large treatment effect was detected using the PLD-Q, reflected by improved quality of life scores. At baseline, the included patients presented with very high symptom scores (PLD-Q: 76.9). By comparison, moderate and severe PLD patients with an equivalent median liver volume (5001 mL) in a previously published cross-sectional study (n=56) had a median PLD-Q of 42.6 (IQR: 30.1 – 48.7).³

In addition to high symptom burden, patients also had a profoundly impaired quality of life (PCS: 24.9, MCS: 40.5). To compare, quality of life of these patients is worse than that of hemodialysis patients (PCS: 33.3, MCS: 47.5) or cirrhotics (PCS: 34, MCS: 44).^{31, 32} After surgery, there were no relevant differences in quality of life between patients and the general reference population for PCS, MCS and EQ-VAS scores, confirming that PHCF achieves a large improvement in this population. Quality of life after PHCF has been previously assessed with SF-36 questionnaire in 69 patients, on average nine years after surgery. Compared to our cohort, PCS was equivalent (45 vs. 45.7) and MCS was higher (50 vs. 55.4) at long-term follow-up.¹²

Three patients (23%) did not achieve clinically relevant improvement after surgery. Our results suggest that the risk of non-response after surgery increases in patients with lower symptoms scores before surgery and with a higher pre-surgical ASA classification. The lower baseline PLD-Q and EQ-VAS score of the non-responders (based on MCID) shows that only highly symptomatic patients should be treated with PHCF, but this cohort is too small to define strict cut-off criteria based on patient-reported outcome measures. We did not see a correlation between change in liver volume and clinical response, but this could be underestimated due to the exclusion of four patients with relatively high baseline liver volume that did not have follow-up imaging.

When looking at individual symptoms of the PLD-Q, only abdominal pain did not significantly improve after surgery. Although it can be caused by a lack of statistical power, it may suggest that abdominal pain in these patients is multifactorial and not only associated with liver volume.

Major complications occurred without mortality in 12% of included patients, which is comparable to previous studies. Two retrospective studies (n=186; n=45) also scored complications after PHCF according to Clavien-Dindo and respectively reported major complications in 21% and 13% of patients and mortality in 3% and 0%.^{13, 33} The relatively higher rate of complications in the first study may be due to general reduction of operative risk since 1985.³⁴ Sensitivity analysis showed that the overall improvement on quality of life was not significantly impacted by the patients with major complications.

An alternative to PHCF in PLD patients with severe hepatomegaly is liver transplantation. In one cohort (n=36), 91% of patients felt better or much better after a mean follow-up time of 62 months.³⁵ Specifically, the symptoms fatigue, physical fitness, loss of appetite and vomiting showed significant improvement.³⁵ The 5-year overall survival (92%) and graft survival (88%) after liver transplantation in PLD patients are high.³⁶ In contrast to symptomatic benefit and good long-term results, perioperative mortality in PLD varies between 4 and 17%, which is higher than in liver transplantation for other indications.^{36, 37} In light of the scarcity of available grafts, the necessity for long-term immunosuppressive medication and high post-operative mortality in a benign disease, PHCF is often preferred.³⁷ Additionally, combined liver-kidney transplantation can be performed in PLD patients with concomitant renal failure.^{38, 39} Although there are some studies with favorable effect of TAE on liver volume, the effect on symptoms and quality of life remains to be elucidated.

The main limitation of this study was the small sample size. This is inevitable for a surgical procedure that is infrequently performed, even in an international referral center. In addition, change in scores could be calculated in 71% to 82% of the cohort despite a high response rate per timepoint. Nonetheless, as the improvement in quality of life after surgery is large, even this small sample size has adequate power to achieve statistically significant and clinically relevant results for our primary and secondary outcomes.

No long-term data on symptom reduction and quality of life is available for our cohort yet. We have chosen a six month follow up as quality of life during postoperative recovery usually normalizes within three to six months after hemihepatectomy.^{40, 41} Although liver volume and symptoms have no linear relationship, we do not expect that quality of life will decline notably over time due to liver growth.³ While recurrence of cysts is common after fenestration, a previous study showed lasting reduction of liver volume after PHCF for up to 20 years.¹³ In addition, equivalent SF-36 scores measured nine years after PHCF, corroborate this.¹²

This study was performed in a referral center that is specialized in treatment of polycystic liver disease, with experienced operating teams. Therefore, the generalizability of these results to other centers may be limited. However, PHCF is an infrequently used and complex surgical procedure due to the altered anatomy and consequently, centralization of this procedure is necessary to achieve the best outcome for the patient.

We calculated the MCID with a distribution-based method based on the standard deviation. Another viable option would have been to calculate the MCID through an anchor-based method. This method uses an additional patient-reported question after the procedure to assess whether the patient thinks a clinically relevant improvement has been achieved.⁴² The PLD-Q does not include a question that could be used to accurately calculate an anchor-based MCID six months after surgery in our cohort. However, as it has been previously shown that half the standard deviation of score changes is equivalent to the MCID, both methods provide an accurate representation of clinically relevant improvement.²⁹

Strengths of this study are that we conducted a prospective study with reliable, responsive and reproducible questionnaires to accurately assess the course of symptoms and quality of life after surgery. In addition, follow-up measurements were performed at predetermined timepoints. Besides statistical significance, we also tested the clinical relevance of our results.

In conclusion, PHCF provides a reduction of symptoms and increase in quality of life in the majority of patients, with values similar to a healthy reference population six months after surgery. However, 23% of patients did not have a clinically relevant improvement of symptoms, despite a decrease in liver volume. In light of potential major complications, careful selection of patients and managing patients’ expectations is important. We confirm that PHCF can be effective for selected patients with moderate or severe PLD and a high symptom burden.

Supplementary Material

1

Supplementary Figure 1: Severity of individual symptoms (%) as scored on the polycystic liver disease questionnaire in the study population before surgery. Scores range from 2 to 11 and are grouped according to severity.

5

Supplementary Figure 2: Radar chart of median individual symptom scores of the polycystic liver disease questionnaire (PLD-Q), pre-surgery (blue) and six months post-surgery (red) for responder and non-responder subgroups.