The Accordion Severity Grading System of surgical complications was introduced in 2009 [24]. Founded on the intent to have an adjustable complication grading system, the Accordion classification provides two versions, using self-explanatory terms rather than grades [24]. The levels range from mild, moderate, and severe complications to the fourth level: death [24]. This grading system, like the Clavien-Dindo classification, is based on the treatment of a complication: thereby, mild complications allow bedside treatment, whereas moderate events need more sophisticated medication (antibiotics, blood transfusions, parenteral nutrition), and severe complications include all interventions as well as organ failure [24]. The third level “severe complications” allows an accordion-like extension for more detailed complication grading. Three categories are offered: “invasive procedures without general anesthesia,” “operation under general anesthesia,” and “organ system failure” [24]. The practicability of the Accordion classification was tested on an international board of experts, and several modifications of the classification were implemented [36, 37]. In addition, the term “sequel of complications” was introduced to grade an advancement inherent to the complication, e.g., the progression of postoperative transient renal failure to chronic persistence, or the occurrence of pyelonephritis after a urinary tract infection [24]. In the attempt to improve this grading system, specifications were added at cost of its simplicity. Possibly due to this deficiency the Accordion Severity Grading System is not widely accepted to date. It is mainly used in Northern America [38, 39] in connection with The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP), a nationwide program in the United States that collects preoperative through 30 days postoperative data on randomly assigned patients [40, 41].

Based on the Accordion Severity Grading System , Strasberg et al. developed an index to quantify surgical complications: the Postoperative Morbidity Index (PMI) [42]. Therefore, complications of the ACS data collection of the NSQIP were graded by the Accordion classification. For each of the six expanded Accordion levels, a numerical weight for severity was calculated: Grade 1, 0.11; Grade 2, 0.26; Grade 3, 0.37; Grade 4, 0.60; Grade 5, 0.79; and Grade 6, 1.00 [42]. To calculate the PMI of one procedure (e.g., laparoscopic colectomy), the severity weights of all complications are summarized and divided by the total number of patients who underwent this procedure [42]. For this calculation only the highest rated complication of each patient is taken into account [42]. However, an external validation on urological procedures found the PMI was insufficient for individual risk adjustment [43]. By considering only the most severe complication, the PMI failed to provide accurate information on patients with more than one postoperative event [43]. So far, the PMI is not widely accepted and was only used for outcome measurements in a few studies outside the authors group [40, 44–46].

To date the Clavien-Dindo classification is the most widely used and accepted postoperative complication assessment classification. Although it allows reporting of all complications, the tabular presentation of multiple complications makes outcome comparison difficult. Therefore, in most studies only the most severe complication is reported. This may lead to underreproting and therfore underestimation of postoperative morbidity [16, 24, 26, 27]. In order to facilitate assessment of patients’ overall postoperative morbidity, the Comprehensive Complication Index (CCI^®) was developed [25]. This complication index is based on the Clavien-Dindo classification [25]. Thereby, every complication’s severity is represented by a number and can be computed with the help of an online calculator, resulting in a number between 0 (no complication) and 100 (death of a patient) [25]. The formula for CCI^® calculation was designed to consider any combination of complications including the ones of lower severity [25]. For its development, methods from operation-risk-index analysis in marketing research were adopted [47–49]. The authors performed an internal and external validation and showed a superiority of the CCI^® over traditionally reported morbidity endpoints, e.g. the Clavien-Dindo classification, simply applicable solely for major complications [25, 50]. With its numeric character, this index easily allows inclusion of all complications for the assessment of the overall postoperative burden. In addition, the comparison of morbidity between clinics and studies is simplified. The CCI^® is internationally increasingly used [51–54] to assess the overall postoperative morbidity since its introduction in 2013 [25].

The most commonly used complication classification systems do not consider intraoperative complications. Of 46 randomized controlled trials published in JAMA Surgery, Annals of Surgery, and BJS in 2010, 41% of the trials failed to report intraoperative complications [55]. To overcome this shortcoming, new grading systems focusing on intraoperative complications [56–59] were developed. Two of them are presented below.

Uniform assessment and reporting of complications are essential for comparing postoperative complications and morbidity [15, 19]. This highlights the importance of complication classification systems, which should ideally be ubiquitary reproducible and consider every complication.

The Clavien-Dindo classification was the first postoperative complication grading system to be based on the treatment of complications and has experienced a widespread acceptance up to this day. Shortcomings of this classification system are the difficulty to compare patients with multiple complications due to its semi-numeric character. For example, the morbidity of a patient with a Grade IVb and a Grade II complication is difficult to compare to the morbidity of another patient with a IIIb and a IVa complication. The CCI^® compensates this shortcoming, by allowing simple calculation of all complications, resulting in a number between 0 and 100. However, it lacks the reflection of what kind of complication a patient endured. For example, in a patient with a Grade IIIb complication, we know he required general anesthesia, whereas, in a patient with a CCI^® of 39, it is unclear if this patient had a Grade IIIb or multiple lower-grade complications. Thus, the CCI^® seems to be a good addition to the Clavien-Dindo classification, rather than a replacement. Intraoperative complications are not recorded in the Clavien-Dindo classification or the CCI^®. Here, novel intraoperative complication classifications such as the CLASSIC seem to be promising.

The CCI^® may also serve as a tool for benchmarking interventional outcomes [54]. The benchmarking concept, known from economic research, implies quality improvement by comparing with the best in class. Self-initiated comparison of postoperative outcome with the best – the benchmark – may reduce postoperative morbidity by reevaluation of attuned processes. In the future, public health decisions may be guided by benchmarks set on the basis of standardized outcome measurements. Financial coverage by health insurances, surgical licensure issued by governmental authorities, or a patient’s choice of hospital may also be strongly influenced by them [54].