Article Text

Download PDFPDF

Predictive risk factors of acute kidney injury after cytoreductive surgery and cisplatin-based hyperthermic intra-peritoneal chemotherapy for ovarian peritoneal carcinomatosis
  1. Martina Aida Angeles1,
  2. François Quenet2,
  3. Pierre Vieille1,3,
  4. Laurence Gladieff4,
  5. Jean Ruiz5,
  6. Muriel Picard5,
  7. Federico Migliorelli6,
  8. Leonor Chaltiel7,
  9. Carlos Martínez-Gómez1,8,
  10. Alejandra Martinez1,8 and
  11. Gwénaël Ferron1,9
  1. 1 Department of Surgical Oncology, Institut Claudius Regaud – Institut Universitaire du Cancer de Toulouse (IUCT) – Oncopole, Toulouse, France
  2. 2 Department of Surgical Oncology, Institut du Cancer de Montpellier (ICM), Montpellier, France
  3. 3 Department of Gynecology, CHR, Perpignan, France
  4. 4 Department of Medical Oncology, Institut Claudius Regaud – Institut Universitaire du Cancer de Toulouse (IUCT) – Oncopole, Toulouse, France
  5. 5 Intensive Care Unit, Institut Claudius Regaud – Institut Universitaire du Cancer de Toulouse (IUCT) – Oncopole, Toulouse, France
  6. 6 Institute Clinic of Gynecology, Obstetrics and Neonatology, Hospital Clínic de Barcelona, Barcelona, Spain
  7. 7 Biostatistics Unit, Institut Claudius Regaud – Institut Universitaire du Cancer de Toulouse (IUCT) – Oncopole, Toulouse, France
  8. 8 INSERM CRCT 1, Toulouse, France
  9. 9 INSERM CRCT 19, Toulouse, France
  1. Correspondence to Gwénaël Ferron, Institut Claudius RegaudInstitut Universitaire du Cancer de Toulouse, Toulouse 31059, France; Ferron.Gwenael{at}iuct-oncopole.fr

Abstract

Objective The aim of our study was to assess the incidence and identify the predictive risk factors of acute kidney injury after cytoreductive surgery and cisplatin-based hyperthermic intra-peritoneal chemotherapy.

Methods This is a retrospective study from two centers evaluating patients with advanced or recurrent ovarian cancer who underwent cytoreductive surgery followed by cisplatin-based hyperthermic intra-peritoneal chemotherapy from January 2007 to December 2013. Patients were classified into two groups according to the occurrence of acute kidney injury, defined as a glomerular filtration rate at post-operative day 7 25% lower than at day 0. We also evaluated acute kidney injury following Risk, Injury, Failure, Lost and End-stage kidney function criteria. Univariate and multivariate analyses were conducted in order to assess the association between different variables and the occurrence of acute kidney injury.

Results Sixty-six patients were included: 29 (44%) underwent first-line treatment and 37 (56%) were treated for recurrent disease. The incidence of post-operative acute kidney injury was 48%. After multivariate analysis, hypertension (OR 18.6; 95% CI 1.9 to 182.3; p=0.012) and low intra-operative diuresis (OR 0.5; 95% CI 0.4 to 0.8; p=0.001) were associated with acute kidney injury.

Conclusion The incidence of acute kidney injury after cytoreductive surgery and cisplatin-based hyperthermic intra-peritoneal chemotherapy was high. Hypertension and low intra-operative diuresis were independent risk factors for this complication. Adequate peri-operative hydration, in order to maintain correct diuresis, could decrease the occurrence of acute kidney injury in patients undergoing cytoreductive surgery plus hyperthermic intra-peritoneal chemotherapy.

  • peritoneal carcinomatosis
  • ovarian cancer
  • hyperthermic intraperitoneal chemotherapy
  • cisplatin
  • RIFLE and nephrotoxicity

Statistics from Altmetric.com

HIGHLIGHTS

  • High blood pressure and low intra-operative diuresis are independent risk factors for acute kidney injury

  • Risk, Injury, Failure, Lost and End-stage kidney function criteria should be employed to define acute kidney injury

  • Standardized peri-operative management of patients is needed in order to reduce the occurrence of acute kidney injury

Introduction

During the last decades, a promising therapy combining cytoreductive surgery plus hyperthermic intra-peritoneal chemotherapy emerged as a new strategy for epithelial ovarian cancer treatment. The addition of hyperthermia to intra-peritoneal drug delivery seems to increase the efficacy of chemotherapy by both a direct cytotoxic and synergistic effect.1 In other rare peritoneal diseases such as peritoneal pseudomyxoma and mesothelioma this strategy has already shown a survival benefit.2 3 Regarding frontline treatment of ovarian cancer, there has been limited data until the recent publication of a multicenter randomized phase III trial, showing a survival advantage in patients undergoing interval cytoreductive surgery with cisplatin-based hyperthermic intra-peritoneal chemotherapy.4 In recurrent ovarian cancer, cytoreductive surgery plus hyperthermic intra-peritoneal chemotherapy has also demonstrated a survival improvement over cytoreductive surgery alone.5

However, hyperthermic intra-peritoneal chemotherapy after debulking surgery is associated with non-negligible morbidity and mortality rates and it has been related to acute kidney injury.6 7 The reported incidence of acute kidney injury is very variable, ranging from 0% to 40%.7–9 Moreover, there is a lack of data regarding the predictive factors of acute kidney injury after cytoreductive surgery plus hyperthermic intra-peritoneal chemotherapy.8–11 We believe that this information would be helpful in order to tailor the peri-operative management of patients undergoing cytoreductive surgery plus cisplatin-based hyperthermic intra-peritoneal chemotherapy and therefore decrease its morbidity.

The aim of our study was to assess the incidence and identify the predictive risk factors of post-operative acute kidney injury in patients with primary advanced or recurrent ovarian cancer undergoing cisplatin-based hyperthermic intra-peritoneal chemotherapy after cytoreductive surgery.

Methods

Patients and study design

A computer-generated search of two institutional patient databases was performed to retrospectively identify all patients who underwent cytoreductive surgery plus cisplatin-based hyperthermic intra-peritoneal chemotherapy for advanced or recurrent ovarian cancer between January 2007 and December 2013 at two Comprehensive Cancer Centers in France: Institut Universitaire du Cancer de Toulouse and Institut du Cancer de Toulouse. Institutional Review Board approval was obtained from both centers.

Surgery

Surgical procedures carried out in both centers were equivalent. Patients with previous anti-hypertensive treatment (angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers) did not receive the treatment on the day of surgery. The surgical technique of cytoreductive surgery was performed following Sugarbaker principles of peritonectomy.12 The surgery was classified into three levels depending on the number of visceral resections and peritonectomy procedures (level I, one or two procedures; level II, three or four procedures; level III, five or more procedures).13 Its main goal was to obtain complete cytoreduction, evaluated using the Completeness Cytoreduction score.14 Debulking surgery was followed by cisplatin-based hyperthermic intra-peritoneal chemotherapy at a dose of 50–100 mg/m2 (decided by the physician in charge), heated at between 41°C and 43°C for 60 min and using an open coliseum technique.15 All patients had previously received platinum-based chemotherapy. Patients in frontline treatment had received a maximum of six cycles while those with recurrent ovarian cancer had received a maximum of 12 cycles. We did not perform hyperthermic intra-peritoneal chemotherapy in patients with pre-operative renal co-morbidity or in those with a physiological age above 70 years.

Study data

The following patient characteristics were retrieved from the clinical records: age, height, weight, co-morbidities, anti-hypertensive treatment, American Society of Anesthesiologists score, previous oncological treatments (chemotherapy or chemotherapy plus surgery), and type of treatment (frontline or recurrence). Pre-operatively administered volume of crystalloids was reviewed. The extent of peritoneal carcinomatosis was assessed intra-operatively using the Sugarbaker Peritoneal Cancer Index.14 The number of visceral resections, level of peritonectomies, Completeness Cytoreduction score, and operative time were also recorded.

Hyperthermic intra-peritoneal chemotherapy-related data were drug dose and body temperature 5–30 min before and after intra-peritoneal chemotherapy perfusion. Intra-operative data collected were epidural analgesia, diuresis, blood loss, administration of red blood cell units, crystalloids and colloids, albumin, aminoglycosides, non-steroidal anti-inflammatories, and continuous intra-venous noradrenaline perfusion. We also collected data on the administration during the first seven post-operative days of red blood cell units, albumin, noradrenaline, aminoglycosides, diuretics, and non-steroidal anti-inflammatories. Results of blood samples performed immediately before surgery, 24 hours after surgery, and at day 7 of hospital stay were retrieved from the records. We searched for post-operative complications.

The main outcome was acute kidney injury, defined as a decrease in glomerular filtration rate of more than 25% at day 7 compared with day 0. The degree of renal failure was classified using Risk, Injury, Failure, Lost and End-stage kidney function criteria for glomerular filtration rate16: Risk, Injury or Failure (glomerular filtration rate decrease >25%,>50% or>75%, respectively). The glomerular filtration rate was calculated using the Modification of Diet in Renal Disease equation.

Statistical analysis

Data were summarized by frequencies and percentages for categorical variables and by medians and ranges for continuous variables. Univariate analysis was performed using χ2 or Fisher’s exact tests for qualitative variables and the Mann–Whitney test for continuous variables. Backward variable selection in a logistic regression model was performed using significant variables (at the 10% level) identified in univariate analysis and those which were considered clinically relevant. The final model was presented with odds ratios (OR) and their 95% confidence intervals (CI). P values <0.05 were considered statistically significant. All statistical analyses were conducted using STATA 13.0 software.

Results

Sixty-six patients meeting the inclusion criteria were identified. The patients’ characteristics are detailed in Table 1. Acute kidney injury was reported in 48.4% of patients (n=30; group A) while 51.6% (n=32; group B) had normal renal function. Data were missing for four patients who were therefore excluded from the analysis. Among the 30 patients with acute kidney injury, according to the Risk, Injury, Failure, Lost and End-stage kidney function classification, 36.7% (11/30) were classified in the Risk group, 33.3% (10/30) in the Injury group and 30% (9/30) in the Failure group. Regarding long-term follow-up of patients with acute kidney injury, data were available for 23/30 individuals (76.7%). Fourteen patients (60.9%) developed chronic kidney disease (defined as glomerular filtration rate <60 mL/min/m2 6 months after surgery) and, among these patients, four (28.6%) required dialysis.

Table 1

Patients’ characteristics

Univariate analysis was performed comparing groups A and B (Table 2). Hypertension and anti-hypertensive treatment were found to be statistically correlated with the occurrence of acute kidney injury. Perfused intra-operative crystalloid volume and diuresis were significantly higher in group B. Although the perfused pre-operative crystalloid volume was also higher in this group, this difference was not statistically significant. Body temperature before hyperthermic intra-peritoneal chemotherapy was significantly higher in group A.

Table 2

Correlation between patients’ characteristics and acute kidney injury (glomerular filtration rate decrease >25% at day 7 after surgery): univariate analysis

The number of patients requiring intra-operative noradrenaline administration was not significantly different between the two groups. However, it must be pointed that, among patients who did not received epidural analgesia, none needed intra-operative noradrenaline perfusion (p<0.001).

Regarding biologic data (Table 3), group A had significantly higher plasma urea levels and a lower glomerular filtration rate than group B at day 1. However, although patients with acute kidney injury seemed to have higher pre-operative urea levels, this difference was not significant (p=0.051).

Table 3

Correlation between biologic data and acute kidney injury (glomerular filtration rate decrease >25% at day 7 after surgery): univariate analysis

No differences were found between the two groups (B vs A) regarding the management during the first seven post-operative days. The median (range) number of red blood cell units administered was 0 in both groups (0–6 vs 0–4, p=0.39). The median (range) number of albumin units (20%, 100 mL) perfused was 4 (0–10) in group B and 5 (0–12) in group A (p=0.93). Only one patient in group B received aminoglycosides (p>0.99). Nine patients (29.0%) in group B received non-steroidal anti-inflammatories compared with 6 (22.2%) in group A (p=0.55). Diuretics were administered to 14 (43.8%) patients in group B and to 16 (57.1%) patients in group A (p=0.30). Four patients (12.5%) received noradrenaline in group B and 5 (18.5%) in group A (p=0.72).

None of the post-operative complications were significantly associated with acute kidney injury and none of the patients needed surgical re-intervention or died during the 30 post-operative days (Table 4).

Table 4

Variable distribution according to RIFLE classification: Risk, Injury, or Failure (GFR decrease > 25%, > 50%, or > 75%, respectively)

Due to the small size of the groups of Risk, Injury, Failure, Lost and End-stage kidney function classification, no comparison was made between them (Table 5). Nevertheless, acute kidney injury severity seems to increase with the reduction of pre-operative and intra-operative volume of perfused crystalloids as well as with low intra-operative diuresis. The proportion of patients who underwent level II or III peritonectomy procedures was 43.7% in patients without acute kidney injury, 54.6% in the Risk group, 70% in the Injury group, and 77.7% in the Failure group. The median cisplatin doses for the four groups were, respectively, 75, 75, 60, and 80 mg/m2.

Table 5

Variable distribution according to the Risk, Injury, Failure, Lost and End-stage kidney function classification: risk, injury or failure (glomerular filtration rate decrease >25%, >50%, or >75%, respectively)

We performed the multivariate analysis including variables that showed a significant association with the outcome in the univariate analysis: high blood pressure, intra-operative crystalloid volume, intra-operative diuresis, body temperature before hyperthermic intra-peritoneal chemotherapy. We also included the following variables, which were considered clinically relevant: pre-operative crystalloid volume, number of peritonectomy procedures, and pre-operative plasma urea level. In our study anti-hypertensive treatment was not included in the multivariate analysis because of its collinearity with the response variable. In fact, none of the patients in the normal renal function group had previous anti-hypertensive treatment. The final model identified two significant risk factors for acute kidney injury: hypertension (OR 18.6, 95% CI 1.9 to 182.3, p=0.012) and intra-operative diuresis (OR 0.54, 95% CI 0.37 to 0.77, p=0.001).

Discussion

The incidence of post-operative acute kidney injury in our study was 48.4%; 30% of these patients were in the Failure group according to the Risk, Injury, Failure, Lost and End-stage kidney function classification. There is a wide range of incidence rates of acute kidney injury reported in the literature. Of the patients described by Kusamura et al, 4% developed renal toxicity grade 3 or higher when using cisplatin±doxorubicin or mitomycin,17 while Bakrin et al observed nephrotoxicity in 8% of patients (with 2% developing chronic renal insufficiency and 1% requiring long-term dialysis).18 Schmidt et al reported only one case of nephrotoxicity among 67 patients when using cisplatin.19 In contrast, other groups have recently published higher incidences of acute kidney injury after hyperthermic intra-peritoneal chemotherapy when following Risk, Injury, Failure, Lost and End-stage kidney function criteria, which is in agreement with our results and could be explained by the strictness of this classification.11 Arjona-Sánchez et al, who used different drugs depending on the origin of the peritoneal malignancy, found acute renal dysfunction in 30.5% of their 141 patients.11 In 2017, Sin et al reported an incidence of post-operative acute kidney injury of 40.4% (19/47) when administering cisplatin, with two patients requiring long-term dialysis.8 The high rate of chronic disease in our study may be due to the lack of registration of the absence of complications, therefore leading to missing data in patients without chronic disease. The exact incidence of acute kidney injury after cytoreductive surgery plus hyperthermic intra-peritoneal chemotherapy is difficult to establish due to the variety of data in the literature. Moreover, authors do not always provide a clear definition of acute kidney injury when reporting its incidence and there is not a general standardized criterion to define it, which could explain why some groups described very low incidences. Arjona-Sánchez et al showed that Risk, Injury, Failure, Lost and End-stage kidney function criteria accurately diagnosed acute kidney injury in patients undergoing hyperthermic intra-peritoneal chemotherapy procedures.11 Thus, we believe that Risk, Injury, Failure, Lost and End-stage kidney function criteria could be used to standardize the definition of post-operative acute kidney injury in patients undergoing cytoreductive surgery plus hyperthermic intra-peritoneal chemotherapy to allow the comparison between different workgroups. However, other classifications could also be used for this purpose, such as Kidney Disease: Improving Global Outcomes, which has been shown to be useful to evaluate acute kidney injury and assess its prognosis in cancer patients treated with cisplatin.20

The univariate analysis showed that hypertension, anti-hypertensive treatment, low intra-operative crystalloid volume, and diuresis were statistically associated with acute kidney injury. In a meta-analysis, Yacoub et al found that pre-operative use of renin angiotensin system blockers was associated with a high risk of post-operative acute kidney injury in patients undergoing cardiovascular surgery.21 Another study, with 390 patients admitted to a medical intensive care unit for septic shock, also showed an association between anti-hypertensive treatment and the development of acute kidney injury.22 Concerning hypertension, a recent study involving cancer patients receiving cisplatin with or without mannitol as nephroprotectant showed that patients with a previous history of hypertension had a higher likelihood of developing nephrotoxicity.23 Another workgroup constructed a predictive model for the occurrence of acute kidney injury after the first course of cisplatin, in which previous history of hypertension was included.24 Analogous to our study, Hakeam et al reported an association of angiotensin II receptor blockers use and hypertension with post-operative acute kidney injury in patients undergoing cytoreductive surgery and hyperthermic intra-peritoneal chemotherapy.10 In our study, hypertension was included in the multivariate analysis, remaining as a significant risk factor. Anti-hypertensive treatment could not be included in the multivariate analysis because none of the patients in the normal renal function group had previous anti-hypertensive treatment. Although it cannot be completely excluded that anti-hypertensive treatment was related to acute kidney injury by itself, it could be hypothesized that hypertension is a predictive risk factor of acute kidney injury due to possible underlying renal injury secondary to hypertension.

We also found that patients who developed acute kidney injury received lower intra-operative crystalloid volume than patients with normal renal function. This last group also had higher intra-operative diuresis. After multivariate analysis, only low diuresis remained as a significant risk factor for post-operative acute kidney injury. These two variables might be related, as a low intra-operative volume of crystalloids could be reflected in a low intra-operative urine output. However, with our data, it was not possible to establish a cut-off of intra-operative diuresis to predict acute kidney injury occurrence. Hakeam et al have already reported an association between low intra-operative diuresis and acute kidney injury development.10 Owusu-Agyemang et al showed that administration of fluid at an average rate of 9 mL/kg/hour was required to maintain a satisfactory urine output in children undergoing cytoreductive surgery plus hyperthermic intra-peritoneal chemotherapy.25 Some reviews showed that hydration remains the most effective strategy to prevent nephrotoxicity in patients treated with cisplatin, as it aids renal clearance of the drug and therefore decreases the amount of time that the drug is in contact with renal tubules.26–28

We did not find either age or body mass index to be associated with acute kidney injury, while other studies had related them with nephrotoxicity.8 9 22 24 In contrast to our work in which the Peritoneal Cancer Index was not significantly associated with post-operative acute kidney injury, Arjona-Sánchez et al related it to acute renal failure after hyperthermic intra-peritoneal chemotherapy.11 Even if not significant, we saw an increased number of peritonectomies in those patients developing more severe nephrotoxicity. In fact, patients with a higher Peritoneal Cancer Index might require more extended peritonectomies, which could lead to more important dehydration and therefore favor acute kidney injury development.

In our study, patients developing acute kidney injury had neither lower baseline glomerular filtration rate nor lower pre-operative albumin levels, while other studies have found these variables to be risk factors of nephrotoxicity.8 22 24 29 Neither administration of red blood cell units nor blood loss were associated with acute kidney injury, as was found by other groups. Peri-operative transfusion is a surrogate marker of intra-operative blood loss, which could lead to hypoperfusion and, consequently, cause renal impairment.8 22 Administration of aminoglycosides and non-steroidal anti-inflammatories also failed to show significance. This is in agreement with the results of Plataki et al, which showed that these drugs were not significantly associated with acute kidney injury.22

The benefit of using inotropes has been controversial. We did not find that noradrenaline administration was either a protective or a risk factor for developing acute kidney injury. Arjona-Sánchez et al identified the use of inotropes as an independent risk factor of acute kidney injury, hypothesizing that it could be related to higher surgical aggression and operative time.11 In contrast, Pili-Floury et al reported that addition of epinephrine to the intra-peritoneal chemotherapy bath reduced both the incidence and severity of nephrotoxicity in patients with ovarian cancer undergoing cytoreductive surgery followed by cisplatin-based intra-peritoneal chemotherapy.

Our results did not show that the cisplatin dose was correlated with acute kidney injury, while other groups have found it to be an independent risk factor for nephrotoxicity.17 23 24 However, the Failure group received the highest median cisplatin dose of all groups. A recently published phase I dose-finding trial of cisplatin-based hyperthermic intra-peritoneal chemotherapy after neoadjuvant chemotherapy and cytoreductive surgery recommends a dose of 70 mg/m2 cisplatin for hyperthermic intra-peritoneal chemotherapy, which is deemed sufficient for maximum efficacy and is considered acceptable in terms of renal toxicity.30 In our study, previous treatment with chemotherapy was not associated with post-operative acute kidney injury. On the other hand, Sin et al showed that both a high number of pre-operative cycles of carboplatin and a short time interval between pre-operative chemotherapy and hyperthermic intra-peritoneal chemotherapy were risk factors for kidney injury.8

The main strength of our study was the assessment of post-operative acute kidney injury using Risk, Injury, Failure, Lost and End-stage kidney function criteria. We believe that its use should be adopted by all workgroups in order to allow comparisons between studies. Moreover, while other studies included patients with different peritoneal malignancies treated with several drugs to assess hyperthermic intra-peritoneal chemotherapy nephrotoxicity, our work was focused on a group of patients with the same diagnosis—ovarian cancer—treated with the same drug (cisplatin). The limitations of this study include its retrospective design, which could lead to possible failure to identify other risk factors of acute kidney injury and inevitably introduce bias. Also, its small sample size might confer lack of statistical power, therefore failing to demonstrate a true association of some variables with the outcome.

The final goal of identifying predictive factors of nephrotoxicity in patients undergoing cytoreductive surgery plus hyperthermic intra-peritoneal chemotherapy is to find strategies in order to diminish this complication (Box 1). As we have seen, adequate hydration is necessary to maintain an acceptable diuresis. Sodium thiosulfate has also been used to prevent cisplatin nephrotoxicity,31 32 but a possible drawback of its use is that it might reduce its desired anti-tumor effect.33 Nevertheless, sodium thiosulfate has been successfully used in a recent phase III trial in which no cases of renal toxicity were reported.4 Likewise, alpha-lipoic acid has also been shown to ameliorate cisplatin-induced acute kidney injury.34 Other strategies to reduce kidney injury could be decreasing the administered drug dose30 or replacing cisplatin by carboplatin, an agent with an improved therapeutic index and lower renal toxicity.35

Box 1

Recommendations to decrease acute kidney injury in patients undergoing cytoreductive surgery followed by hyperthermic intra-peritoneal chemotherapy

General recommendations

  • Ensure adequate peri-operative hydration to maintain correct diuresis

  • Using sodium thiosulfate or alpha-lipoic acid as nephroprotectants should be considered

  • Patients with hypertension should be managed with special care as they are at higher risk of developing acute kidney injury

  • Anti-hypertensive treatment should not be administered on the day of surgery

Conclusion

The incidence of acute kidney injury after cytoreductive surgery followed by hyperthermic intra-peritoneal chemotherapy with cisplatin is high. Hypertension and low intra-operative diuresis were identified as predictive risk factors of acute kidney injury development. Larger prospective multicenter studies should be performed to establish a standardized protocol of peri-operative management for patients undergoing cytoreductive surgery plus hyperthermic intra-peritoneal chemotherapy in order to minimize acute kidney injury.

Acknowledgments

MAA acknowledges the grant support from ”la Caixa” Foundation, Barcelona (Spain). CM-G acknowledges the grant support from Alfonso Martín Escudero Foundation, Madrid (Spain). The authors acknowledge Philippe Izard, Etienne Chatelut, Juan José Torrent, Sébastien Carrere, and Olivia Sgarbura for their contribution to the manuscript.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.

Footnotes

  • Contributors MAA: Conceptualization, data curation, methodology, writing, original draft. FQ: Conceptualization, project administration, methodology writing, review. PV: Conceptualization, data curation, methodology, writing, original draft. LG: Conceptualization, project administration, methodology writing, review. JR: Conceptualization, project administration, methodology writing, review. MP: Conceptualization, project administration, methodology writing, review. FM: Conceptualization, data curation, methodology, writing, original draft. LC: Methodology, formal analysis. CM: Data curation, methodology, writing, original draft. AM: Conceptualization, project administration, methodology, writing, review. GF: Conceptualization, project administration, methodology, writing, review.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned, externally peer reviewed.

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.