The H.O.U.S.E. classification: a novel endoscopic retrograde cholangiopancreatography (ERCP) complexity grading scale

Endoscopic retrograde cholangiopancreatography (ERCP) is a technically challenging endoscopic procedure, where severe complications cannot be avoided [1‐3]. Risk factors behind the development of complications are; the experience of the endoscopist [4‐11], case volume of procedures [4, 5, 7‐12], the indication for the procedure [7‐9] as well as the technical complexity [7‐9]. Additional factors affecting outcome are those associated with altered anatomy after gastric surgery, while others are more unpredictable, like a giant duodenal diverticula [13, 14] or a small papilla requiring a precut sphincterotomy for access. Hence some risk factors cannot be anticipated but the majority can be defined in advance and in the ideal world there is room available for prophylactic, preventive measures to be taken to minimize the risk profile for each individual patient. Classifications of the complexity of ERCP have been presented [15, 16], but do not include modern endoscopic treatment modalities. Cotton and coworkers [6] presented a classification system where experienced endoscopists were asked to grade different endoscopic procedures according to their complexity to which was added also their own experience. A median value of their scores was thereafter filed, building up a grading system for the complexity of the ERCPs. Accordingly, this scale represents the only complexity grading scale established during the last decade but it is entirely “eminence-based”. Moreover, available scoring systems in the field of endoscopic complexity grading [6, 15, 16] have never been validated and have also become outdated due to introduction of new diagnostic and treatment facilities. This is of course problematic in itself, but the current situation also hampers the communication of outcomes in clinical practice but obviously so regarding the presentation of research outcomes. There are several additional reasons for introducing a new complexity grading scale in ERCP. Such a grading system has the potential to function as an aid in an ERCP-education program, to find the right level in a step-wise training curricula [15]. It may facilitate the planning and allocation of resources and the amount of time required for a certain procedure. It may also offer a solid background for adequate charging of endoscopic procedures between different healthcare systems and providers. Finally, it may be critically important to assist in defining the right level of competence and experience of the individual endoscopist in patients who are planned to be referred to a tertiary referral-center.

The objectives of the present study were therefore to define, test and validate a new classification system for ERCP with the ambition to address the defined unmet needs of today and in the foreseeable future.

Two thousand one hundred eighty-five ERCP procedures at Karolinska University Hospital, Huddinge were registered in GallRiks between 2009 and 2011. Of these, 1954 procedures were index-ERCPs, the first ERCP in every treatment episode, and accordingly included in the subsequent analyses. In 23 cases complete 30-day follow-up data were not captured why these were excluded and resulting in 1931 ERCP procedures being analyzed (Fig. 1). The medical records of all remaining patients were reviewed and the intra- as well as post-procedural complications and cannulation success rates were recorded from the GallRiks Registry into every single HOUSE class (HOUSE 1, n = 1124; HOUSE 2, n = 541; HOUSE 3, n = 266). Demographics and indications of the respective HOUSE classes are given in Table 2. The pre-procedural pancreatitis rate in HOUSE class 3 was significantly lower than that of HOUSE class 2 (P = 0.0235) but not compared to HOUSE class 1.

The mean procedure time in HOUSE 1 was (40 ± 0.7 min) whereas the procedure times increased significantly in HOUSE 2 (65 ± 1.5 min) and HOUSE 3 (106 ± 3.2 min), respectively (Table 3). The correlation between the ERCP procedure times (min) and the respective classes of the two different complexity grading systems (HOUSE, Cotton) are given in Fig. 2. However, bile duct cannulation rates were significantly higher in the HOUSE 2 and 3 groups probably because previous sphincterotomy had been done to a much larger extent in these two groups. More advanced procedures had also, by definition, been done especially in the HOUSE 3 group (Table 3 and Fig. 3).

The overall intraoperative adverse event rate was 4.2% and 244 of the 1931 patients (12.6%) experienced postoperative complications. The overall post ERCP pancreatitis (PEP) rate was 4.9% and the incidence of postoperative bleeding was 2.5% (Table 4). The overall postoperative adverse event rate, at the 30-day follow-up, was 11.1%. in HOUSE class 1. With increasing HOUSE-classifications we recorded a significant increase in the postoperative adverse event rates, as illustrated by 15.7 and 12.8% in HOUSE 2 and 3, respectively (Table 4). The PEP incidence was 3.4% in HOUSE class 1 which increased significantly to 7.0 and 6.8% in HOUSE 2 and 3, respectively (Table 4). The pancreatitis frequency by the respective classes, as scored by the two different complexity grading systems (HOUSE, Cotton), are given in Fig. 4. We were unable to detect any differences between the HOUSE groups regarding sphincterotomy bleeding risk nor in abscess formation rates (Table 4).

The main finding concerning this new HOUSE-classification is that it co-varies well to the complexity of the ERCP procedures as well as to procedure time and, to some extent, to adverse events. The HOUSE-classification is also applicable to and in line with modern endoscopic treatment procedures in ERCP. This indicates that the HOUSE-classification could be used in clinical practice to optimize resource planning for each individual ERCP procedure.

In the complexity-grading scale developed by Schutz [16], no differences in complication rates were seen relating to the complexity grading of the ERCP, except for the success-rates of the procedures. Since radiological and endoscopic techniques have developed after the introduction and establishment of this grading scale, to the continued implementation of these criteria in modern endoscopic treatment is problematic, e.g. a categorization system has to incorporate procedures regardless of whether those are diagnostic or therapeutic. This represents an important limitation since the diagnostic procedures to a large extent have been abandoned, after the introduction of the MRI-technique. The Morriston Hospital Grading Scale by Ragunath et al. [15], was mainly developed for educational purposes. More importantly these investigators were unable to demonstrate significant differences in complication rates between the different complexity grades. The complexity grading system, as launched by Cotton et al. [6], where the classification system was presented based on experienced endoscopists’ personal opinion concerning the complexity of different endoscopic procedures. The calculated median values of these experts’ opinion scores, created the grading system. Accordingly, this was not generated out of a clinical prospective validation process but entirely on eminence-based opinions. Hereby we present a classification system that can be used in clinical daily practice and which correlates to the duration of each individual procedure and post-interventional complications. The calculations and estimates were constructed on a critical review of the clinical courses of close to 2000 ERCP procedures at the Karolinska University Hospital, carried out during a defined timeperiod.

There are several arguments in favor of the introduction of a validated complexity classification system. One of these is to target and allocate adequate resources to where they are needed, e.g. to schedule the right amount of time in the endoscopy suite and/or the appropriate competence and equipment necessary for a specified procedure. Another argument in favor of a corresponding classification can be that it fosters and endorses education and training. Hereby both the trainee and the supervisor could select the right level of procedure complexity, adjusted to the trainee’s current skills and educational program. A third argument is that such a classification system can offer a relevant and precise tool to guide the more complex ERCP-procedures directly to the tertiary referral-centers. Such a grading system may also simplify and enhance the accuracy of the billings of endoscopic procedures harmonizing between different health care systems and providers. A new classification would facilitate the option for intra- and inter-individual comparisons within and between different endoscopic centers regarding case mix, success- and complication rates in a more objective and constructive way. Last but not least clinical research obviously requires well validated instruments to be used when introducing various means through which outcomes can be more safe and effective.

However, there are apparent limitations confined to the study design in the form of selection bias. The study was performed in a tertiary referral center where more complex procedures are over-represented compared with many other endoscopy units. Furthermore, one important driving force behind the development of the classification scoring system was to gain control of the increasing costs for endoscopic procedures. Thus, some of the variables contained within the score are not apparent prior to the actual examination (e.g. if multiple stents were introduced or if there is a “caged” papilla). With the further validation and implementation of the HOUSE classification it might well have to be modified to make it even more clinically relevant and it has also to be tested in other clinical environments, as represented by the non-referral centers [4, 10‐12]. Another aspect that has to be taken into consideration is that Karolinska University Hospital is a teaching hospital in endoscopy and it is well known that in an ERCP training situation there is an increased risk for complications [8], diluting the success rates and prolonging the actual procedure times. These aspects warrant the attention of further studies, possible additional modifications and validation of the HOUSE classification in a non-educational endoscopic center.

The HOUSE classification appears to be easy to implement in clinical praxis and facilitates the planning of procedure time and the resources needed for the endoscopy room. It harbors the potential to facilitate education and the endoscopist’s individual training curriculum. However, since the HOUSE-score, in its present form, contains some peri-procedural criteria that cannot be predicted in advance it cannot be used as an objective instrument to direct more complex endoscopic procedures towards tertiary referral-centers.

In this article a new complexity grading scale for ERCP (HOUSE-classification) is described that can be used in daily clinical practice to optimize resources and allow the comparisons of results between different endoscopic centers. However, further studies are warranted to further sharpen this instruments validitity and general clinical relevance.