Predictors of gallstone composition in 1025 symptomatic gallstones from Northern Germany

Gallstones represent a serious burden for Western healthcare systems: 10–20% of Europeans and Americans carry gallbladder stones [1, 2]. The prevalence of gallstone disease is rising, possibly as a result of longer life expectancy and altered nutritional habits. Many gallstones are silent, but symptoms and complications ensue in around 25–50% of cases, necessitating surgical removal of the gallbladder, usually by laparoscopic cholecystectomy [2, 3]. Each year, more than 170,000 cholecystectomies are performed in Germany [4]. Cholelithiasis incurs annual medical expenses in excess of $6 billion in the US and is currently the second most expensive digestive disease, exceeded only by reflux disease [5]. The clinical management of gallstone disease is almost exclusively based on cholecystectomy and endoscopic or medical treatment of complications. Cholecystectomy, although an established procedure, still carries a small but existent complication rate, especially when performed in an acute setting. Mortality rates following cholecystectomy range from less than 0.1% in clinical studies to 0.7% (as documented for all cholecystectomies performed in Germany in 2004) [4]. In the US, about 3,000 deaths (0.12% of all deaths) per year are attributed to complications of cholelithiasis and gallbladder disease [6].

An improved etiological and pathophysiological understanding of gallstone disease may lead to novel, non-surgical, options for prevention and therapy. Clinically, high risk groups with small gallstones have recently been defined [7]. Significant progress has been made both in the genetics of gallstone formation [1, 8] and in the molecular biology of bile excretion [9–11]. Consequently, this mechanistic knowledge may be lead to novel non-surgical therapeutic or preventive strategies, as for instance shown by the prevention of cholesterol gallstone formation in a mouse model by FXR agonists [12] or the cholesterol absorption inhibitor ezetimibe [13]. Gallstones, however, are heteroneous both in terms of their composition and their pathogenesis. The application of novel preventive or therapeutic approaches will likely be limited to certain classes of gallstones as defined by their composition and etiology. Therefore, in order to select the correct patient groups for such interventions, there is a need to idenitify clinical predictors of stone composition.

Gallstone disease is a disorder with a changing prevalence, reflecting the increasing life expectancy and changes in life style in "westernized" societies [1]. In particular, the increase of life style related risk factors was assumed to result primarily in an increase of cholesterol gallstones [14]. Gallstone composition, however, has not attracted adequate attention in current population-based studies leaving uncertainty regarding the confirmation of this assumption. The impact of life style and ethnicity is underscored by the profoundly different gallstone compositions in recent studies from sub-Saharan Africa [15] and China [16, 17].

Classical chemical analysis of gallstone samples is a very laborious methodology. Fourier Transform Infrared (FTIR) spectroscopy has been established as a means of gallstone analysis by Japanese [18] and US investigators [19]. The quantitative reliability has been confirmed in systematic admixtures studies [19]. Thus, the utilization of FTIR makes the analysis of larger series of gallstones samples possible. We thus used FTIR, in order to define target populations for future therapeutic and preventive therapies and to investigate the possibly changing gallstone composition on a background of the rising prevalence. We obtained gallstone specimens from 1074 cholecystectomies from Northern Germany. We generated descriptive measures of stone composition in this population and investigated potential predictors of stone composition.

Fourier Transform Infrared (FTIR) spectroscopy has been established as a means to the rapid assessment of the composition of gallstones [18, 19]. Only this methodology enabled the analysis of the over 1000 stone samples and was thus used to investigate this large sample of gallstones. We have chosen to present the results of the FTIR in a semi-quantitative classification rather than exact numbers of the proportion of the respective stone components, because the direct quantification of components using FTIR is limited by some interfering factors: Namely, interaction of the components with the KBr, differences in hydration, ion exchange reactions, the degree of dispersion between KBr and gallstone components, the Christiansen-effect and/or possible chemical reactions of components can influence the relative height and appearance of the spectrum peaks [22–24]. Thus, in order to obtain robust (and conservative) estimates of gallstone composition, the substances were classified into "main", "intermediate" and "trace" components (as described in the Methods section) rather than being expressed in explicit percentages.

The fine structure [21, 25], bacterial colonization [26] and mechanistics of gallstone generation [27–30] have been studied in detail in previous experiments. We think, that this study still makes a substantial contribution in that it has focussed more on a general epidemiological survey of gallstone composition mainly to determine the principal components of the current gallstones in order to help identify patients for future innovative non-surgical interventions [12]: In the descriptive statistics of gallstone composition, two main conclusions can be drawn: Firstly, gallstones in the investigated population are quite homogenous in terms of their main components: In only eightytwo (7.6%, including sixty stones, that contained bilirubin/cholesterol combinations) out of 1074 stones, different main substances were identified in the two samples taken from each stone (Table 3). Most of this relates to the inclusion of bilirubin in otherwise dominant cholesterol stones, thus representing the pigment cores or rings in cholesterol gallstones, which have been studied previously in detail [21]. The other source of heterogeneity stems from calcium as an admixture to gallstones [25, 31]. The quantitatively major components of gallstones identified in this experiments are cholesterol, bilirubin (pigment) [32] and calcium carbonate [31] with 93.3%, 5.5% and 4.8% relative frequencies of detection as main substances (Table 2). These major components are not randomly assembled in gallstones (p < 0.001), but the composition falls into distinct classes of gallstones, namely cholesterol, bilirubin (pigment) and calcium stones. This classification has obviously been developed before and is here confirmed on a more formal level. Thus, for further analysis, stones were classified into four classes (for exact definition see the Results section). Calcium was detected only in 10% of gallstones, which has implication for the sonographic and radiographic detection of these concrements.

Interestingly, the age of onset was not strongly correlated to stone composition (p = 0.045), this is probably due to the lack of coverage of the paediatric population in this survey: The youngest patients in this study were 18 years at diagnosis. Thus, the reported very different gallstone composition of approximately 50% pigment stones and 35% calcium stones in children [33, 34] was not observed here.

As a main result of the search for epidemiological predictors of stone composition, sex and BMI were identified: Both associations were significant (p < 0.001). Interestingly, in obese individuals (BMI≥30), approximately 95% of gallstones consist predominantly of cholesterol for both men and women. Thus, the increase in gallstone prevalence and the association of overall gallstone risk with higher BMI [2] are likely resulting primarily in an increase of cholesterol gallstones.

The study has addressed a selected patient population and is thus not population-representative: This applies to the focus on younger patients who were operated at an age under 65, the overall response rate as compared to the total operated population of (10.3%) and to potential bias introduced through storage of gallstone samples by the patients. Here, the personality traits of the patients (orderliness etc.) and the stability of stones may have played a role. Stones that tend to disintegrate are more likely to be of rarer composition types (polysaccharide, certain calcium stones) and may have been thus underrepresented in this investigation. The overall pattern of stone composition was similar in the 87 consecutive patients directly recruited at the Kiel university hospital in 2005 thus indicating, that main findings may robust against these potential selection bias factors.

Taken the limitations noted above into account, the study shows the changing patterns of gallstone disease and confirms the need for the present investigation. In this experiment cholesterol stones represent the by far dominating type of gallstone. In contrast, studies from the 1960is and 1970ies have shown prevalences of pigment stones of 23 [19] to 30 percent [35]. In our study, pigment stones are relatively infrequent with only two percent of stones in the present sample of symptomatic gallstones. Previous studies from Western countries have also shown a predominance of cholesterol stones with 100% of patients under the age of 50 and 60% for older patients [36] in a Dutch study and 58% of cholesterol stones in a U.S. study using visual inspection of stones [19]. If the definition of a cholesterol stone from van Erpecum et al. is adapted to our data, 91.1% of stones contain 70% or more cholesterol (row 1 of Table 2). Some 2.2% of stones contain between 30 and 70% cholesterol and cannot be clearly by put into the 50% categorisation (rows 2–3 of Table 2). Although no correlation of age of operation and stone composition was found, this might be due to the exclusion of old patients with gallstones. Thus, the high proportion of cholesterol gallstones in this investigation may be due to this age selection and also related to the same altered nutritional patterns that also lead to an increase in gallstone prevalence in susceptible populations.

In summary, this study documents a dramatic change in gallstone composition – in the high-prevalence region of Northern Germany – with a overwhelming preponderance of cholesterol-based gallstones. This is in contrast to the gallstone composition in recent studies from sub-Saharan Africa [15] and China [16, 17]. Based on the findings of this study, a potential homogenous target population in Germany for non-surgical intervention [12, 13] for the prevention of cholesterol stones would be obese individuals with a BMI > 30, for which approximately 95% of cholesterol gallstones were observed in both males and females.