Urinary saturation: casual or causal risk factor in urolithiasis?


  • Allen L. Rodgers

    Corresponding author
    1. Department of Chemistry, University of Cape Town, Cape Town, South Africa
    • Correspondence: Allen L. Rodgers, Department of Chemistry, University of Cape Town, Rondebosch, Cape Town 7701, South Africa.

      e-mail: allen.rodgers@uct.ac.za

    Search for more papers by this author



  • To assess (i) the extent to which urinary supersaturation (SS) has successfully discriminated between stone formers and healthy individuals (N), (ii) whether absolute SS has diagnostic worth and (iii) whether high SS is the fundamental cause of stone formation per se.

Materials and Methods

  • Google Scholar was used to identify studies in which urinary compositional data had been determined.
  • In those cases where SS values were not given, or where other risk indices had been reported, they were (re-)calculated.
  • Collected data were termed ‘global’ but were then ‘filtered’ according to stone type and protocols used for SS calculations.
  • SS distribution plots for calcium oxalate, brushite and uric acid were constructed.
  • Data were statistically analysed using the unpaired t-test and Mann–Whitney test.


  • In all, 47 studies yielded 123 SS values for healthy individuals and 122 values for stone formers.
  • The mean and median SS values were significantly greater in stone formers compared with healthy individuals in all but one of the comparisons.
  • Wide variations in SS occurred for healthy individuals and stone formers. The two groups could not be separated.


  • Absolute SS has no diagnostic worth. It is impossible to quantify the meaning of a ‘high’ SS value. Urines cannot be identified as originating from healthy individuals or stone formers based on their SS.
  • SS should be determined in clinical and research settings for relative comparisons during the assessment of treatment efficacies.
  • This study provides a compelling argument for SS being a casual factor rather than a causal one.


The saturation of urinary salts has been cited for >40 years as a key risk factor for urolithiasis. Robertson et al. [1-3] were among the first to invoke the concept of saturation in stone research, and to show how it should be calculated, in a series of papers published in the 1960s. The capacity of stone researchers to calculate this apparently crucial parameter was boosted dramatically with the launch of the software package EQUIL [4] and its updated versions EQUIL2 [5] and EQUIL 93 [6]. It would not be an exaggeration to suggest that the vast majority of studies investigating urinary lithogenic risk factors have used these computer programs. Other programs such as SUPERSAT [3], SEQUIL [7], URSUS [8] (cited by Trinchieri et al. [9]), MINEQL [10] and JESS [11, 12] have also been used. However, irrespective of the computer program, authors have unwaveringly based their studies on the premise that the magnitude of the urinary saturation is a measure of the risk of stone formation.

Despite its extensive citation in the literature as a risk factor for stone formation, the terminology used to describe the urinary state of saturation has varied widely, even when it has been generated by the same software package. The output of EQUIL has been referred to as activity product, activity product ratio, saturation, relative saturation, supersaturation, and as the difference in the Gibbs free energy (ΔG). The use of these terms is potentially confusing, given that there are differences between some of them. It might be helpful to clarify these here.

According to Kavanagh [13, 14], for practical purposes, the supersaturation ratio ‘S’ for a salt such as calcium oxalate (CaOx) is given by: S = [Ca2+][Ox2–]/[Ca2+ eq][Ox2–eq], where the square brackets represent molar concentrations, the numerator represents the ion concentration product (activity product) in a given solution and the denominator represents the ion concentration product corresponding to an equilibrium state of saturation (solubility product).

EQUIL2 generates two different outputs, ‘S’ and ΔG [5]. The present writer believes that authors who have used the terms ‘activity product ratio’, ‘saturation’ and ‘supersaturation’ are actually referring to the supersaturation ratio as defined above. Use of these terms is perfectly acceptable. However, the term ‘relative saturation’ (or ‘relative supersaturation’) has a different meaning. It is given by the expression σ = S–1. Its synonymous use with ‘S’ in several studies, (all of which employed EQUIL software to generate its values), has therefore been inappropriate. Similarly, the term ‘activity product’ is not equivalent to the others; it refers to the numerator in the expression for ‘S’ given above. Although it has not been synonymously used with the latter, care needs to be exercised in recognising that it is different to S per se. Finally, the differential Gibbs free energy is a measure of the thermodynamic pressure for crystals to form [5]. It is not numerically equal to ‘S’, so their respective values should not be compared. Nevertheless, it provides a measure of the probability or risk of crystallisation occurring in a particular urine. The supersaturation ratio as defined above is also the output of the program JESS, which has been recently used in urolithiasis research [15], but it is referred to as the saturation index (SI) in this software package [11]. The terms ‘S’ and ‘SI’ are therefore equivalent.

Since urinary saturation has been invoked as an all-important risk factor for stone formation for the past four-to-five decades (irrespective of the terms that have been used to describe it), the present author undertook to investigate the extent to which it has been a successful discriminator between stone formers and healthy individuals and whether relatively high values are the fundamental cause of the disease per se. For the purposes of the present paper, the term ‘supersaturation’ and the symbol ‘SS’ will be used to represent the supersaturation ratio.

Materials and Methods

The search engine Google Scholar was used to identify studies in which urine compositional parameters had been determined. The keywords ‘urolithiasis risk factors’, ‘urinary lithogenic risk factors’, ‘urine composition’, ‘urinary saturation’, ‘urolithiasis trials’ and various combinations thereof were used for the search. Inclusion criteria were studies in which healthy individuals or stone formers or both had been investigated, healthy individuals and stone formers had been considered independently, subjects had collected 24-h urine samples while on their free unrestricted diets, and urine compositional data were reported. Only baseline urine values were included in the study.

SS values for calcium oxalate monohydrate (COM), brushite (Bru) and uric acid (UA) in healthy individuals and stone formers were collated from the literature. In those cases where SS values were not given in the published article, or where risk indices other than SS had been reported (such as the Tiselius risk index or the activity product), they were calculated with EQUIL2 by the present author to render them compatible with values that had been similarly calculated in other studies, using the reported mean urine composition values. Similarly, in those cases where ΔG was reported in the original paper, it was converted to SS. Finally, SS values which appeared to be suspiciously high or low, were checked by recalculation (again using the published urinary data and EQUIL2) and were replaced with the corrected values whenever necessary.

In those cases where the distributions of SS values for a particular salt in healthy individuals and stone formers were Gaussian, the unpaired t-test was used to determine whether the respective means were significantly different or not. The Mann–Whitney test on the medians was used for distributions that failed the normality test.


In all, 47 studies which satisfied the inclusion criteria were identified. Among these, there were 14 studies in which stone formers and healthy individuals were each investigated, while there were 18 studies involving only healthy individuals and 15 studies involving only stone formers. Besides the variation in the terminology used to describe SS and the software used in its calculation, several other inter-study inconsistencies emerged with respect to the generation of this parameter. In some studies, patients having different stone types were grouped together; in other studies stone type was not reported. There also were differences in the number of urinary components used to calculate SS. Of the 47 studies surveyed for the present paper, 16 used 13 parameters (calcium, magnesium, sodium, potassium, oxalate, citrate, phosphate, sulphate, chloride, UA, ammonium, pH and volume), 26 used 10–12 parameters while four used less than nine parameters. Finally, while several studies distinguished between male and female subjects, many included participants of both genders in the same study group without making any distinction between them.

The 47 studies yielded 46 values for SS(COM), 39 for SS(Bru) and 38 for SS(UA) in healthy individuals, and 48, 40 and 34 values respectively in stone formers. Of these 245 SS values, 77 had to be calculated by the present author for the above-mentioned reasons, while three values reported in the original studies had to be corrected. Because of the heterogeneous nature of all the data, they will henceforth be referred to as the ‘global’ database. Their respective distributions are shown in Fig. 1A–C.

Figure 1.

Distribution of SS values in the global database: A, SS(COM); B, SS(Bru) and C, SS(UA). N, healthy individuals; SF, stone formers; horizontal bars express two standard errors above and below the mean.

In an attempt to achieve a semblance of homogeneity for inter-study comparisons, the global database was then ‘filtered’ with respect to patient type, software package used for calculating SS values and the number of urinary components used as input for the latter. Accordingly, inclusion criteria for the next set of analyses were defined as idiopathic calcium stone formers, EQUIL2 and the aforementioned list of 13 parameters, respectively. Only those SS values which were derived according to these criteria were used for the additional analysis. The filtered SS values were then re-plotted (Fig. 2A–C).

Figure 2.

Distribution of SS values in the filtered database: A, SS(COM); B, SS(Bru) and C, SS(UA). N, healthy individuals; SF, stone formers; Horizontal bars express two standard errors above and below the mean.

The range of SS values in healthy individuals and stone formers for different urinary stone components for global and filtered databases is given in Table 1. The mean and median SS values for the global and filtered databases are given in Table 2. The upper and lower SS limits of the overlap regions (‘grey zones’) in healthy individuals and stone formers for COM, Bru and UA as well as the percentage of SS values lying in each of these zones is given in Table 3.

Table 1. Range of SS values in healthy individuals and stone formers for different urinary stone components for the global and filtered databases
Healthy individualsStone formersHealthy individualsStone formers
Table 2. Mean (sem; median) SS values in healthy individuals and stone formers for the global and filtered databases
Healthy individualsStone formersPHealthy individualsStone formersP
  1. *Non-Gaussian distribution.
SS(COM)4.25 (0.29; 4.45)6.96 (0.36; 7.17)<0.0013.33 (0.54; 3.03)7.95 (0.65; 7.86)<0.001
SS(Bru)1.04 (0.09; 0.86)*1.27 (0.07; 1.24)*0.0110.71 (0.09; 0.75)1.32 (0.09; 1.28)<0.001
SS(UA)1.33 (0.68; 1.25)*1.67 (0.76; 1.51)*0.0371.29 (0.25; 1.13)1.83 (0.33; 1.64)0.204
Table 3. Range and extent of overlap of SS values for healthy individuals and stone formers in the global and filtered databases
  Grey zone range% of SS values lying in grey zone
SS(COM)Global (Fig. 1A)2.37–8.4073.4
Filtered (Fig. 2A)4.28–6.3828.0
SS(Bru)Global (Fig. 1B)0.16–2.0097.5
Filtered (Fig. 2B)0.74–1.2852.2
SS(UA)Global (Fig. 1C)0.56–3.2593.1
Filtered (Fig. 2C)0.66–2.6575.0


Despite the central role that SS has played in numerous studies for the assessment of the risk of stone formation, the present study has shown that there have been widespread inconsistencies in the calculation and in the reporting of this apparently important parameter. Consequently, there is wide variation in the values of SS for healthy individuals and stone formers (Table 1), with significantly large overlap (grey zones) of their values (Table 3). This is particularly evident in the global plots for SS(Bru) (Fig. 1B) and SS(UA) (Fig. 1C), where the overlap in both is >90% (Table 3). This is hardly surprising in the global studies, given the extremely heterogeneous nature of the data used to generate the SS values. Besides having been derived by different software packages, the actual number of input parameters has varied, as have the clinical pathologies of the patients from whom they were obtained. Filtration of the data improves the situation by decreasing the range of the grey zones (Table 3, Fig. 2A–C). However, despite the improved separation of the SS values in healthy individuals and stone formers for the three urinary salts, overlap still occurs to a disturbing extent. Even in the best case scenario, SS(COM) filtered, the overlap is 28% (Table 3). Clearly, neither of the databases (global or filtered) allows healthy individuals and stone formers to be convincingly separated.

Further discrimination between SS values for healthy individuals and stone formers might be achieved by introducing more ‘filters’ in the hope of reducing the range of the grey zones. An obvious filter would be that of gender. It is somewhat surprising that many investigators have included subjects of both genders in their study groups, seemingly disregarding the evidence that strongly supports the notion that males and females have different stone-risk profiles. Several epidemiological studies have shown that the prevalence of renal stones is lower in females than in males [16, 17]. Basic science studies comparing urine chemistries in males and females have also shown differences. For example, Hesse et al. [18] investigated healthy subjects from both genders in six different age groups and found that SS(COM) values for males were significantly greater than those for females in three of these groups. Asplin et al. [19] have also reported that SS differences occur between normal males and females but they comment that such differences are minimal in stone patients. In the present survey, it was not possible to construct meaningful SS distribution plots for male vs female for ‘normal’ and stone-forming urines as there was only one normal female study and only four normal male studies, which satisfied the strict inclusion criteria described earlier in this paper, and only two stone formers studies in which males and females were investigated separately. Moreover, the older version of EQUIL was used in the latter two studies, which would have precluded them from the analysis anyway. Thus, the present survey does not permit the writer to make a definitive judgement on this issue; however, it is suggested that a gender filter should be considered in future studies of this nature.

Another confounding factor in efforts to achieve homogeneous data sets is the timing of the 24-h urine collections. Clearly, these will not necessarily coincide with high-risk periods for stone formation and will vary widely from one study to another. Although standardisation of this factor would contribute towards filtration of the data for comparative purposes, it would be difficult to achieve

The results of the present study show that there is little, if any, diagnostic worth in an individual's SS value, unless it lies below the lower, or above the upper thresholds of the respective grey zones (Table 3), in which case a probability-based prediction can be tentatively made. Alarmingly, any SS values lying between these extremes could be equally classified as originating from a non-stone former or from a patient with nephrolithiasis. Even when SS is determined using the filtered database, the best case scenario involving SS(COM) shows that 28% of individuals may be misclassified. Thus, given a particular urinary SS value, it is impossible to identify whether the urine is from a stone former or healthy individual. The unreliability of SS as a diagnostic measure is even more convincingly illustrated by inconsistencies that occur in the same laboratory, notwithstanding that the protocols for their determination are identical. Thus, Borghi et al. [20] found SS(COM) values of 10.1 and 11.2 for two groups of idiopathic calcium stone formers, yet 2 years later, a similar clinical group of patients had a mean value of 5.7 [21], representing a difference of nearly 50%. Similarly, Siener et al. [22, 23] found SS(COM) values of 4.28 and 7.20 in two groups of recurrent CaOx stone formers who were respectively investigated 2 years apart. Dismissing this argument on the basis of the fact that different groups of patients were involved within each laboratory (albeit that they had the same clinical profiles) is effectively conceding that absolute SS values have little worth in a clinical or research setting. Indeed, one may query the clinical worth of knowing a patient's SS value in the first place. The present writer believes that it's only worth is to provide a basis for a relative comparison with values derived during and after treatment. Striving for a decrease in SS has been standard practise ever since the concept of SS was first invoked as a central feature of stone formation, and it makes sense, as the requirement of a supersaturated urine for crystallization is undisputed [4, 13, 24]. A decrease in SS, typically achieved by a high fluid intake in a clinical setting, can therefore be appropriately interpreted as reducing the risk of stone formation, but the core question still remains whether high SS is the cause of stone formation per se.

In this context, what does ‘high’ mean? Should the lower and upper thresholds of the respective grey zones (Table 3) be used as the standards on which to judge ‘high’ and ‘low’ values or should the averages (means and medians) derived in the present review (Table 2) rather be used? Recognising that the latter are merely measures of centrality, they can at best be regarded only as very approximate guides as to what might constitute ‘safe’ and ‘perilous’ SS values. The threshold values are more promising in this regard, but are also tenuous. Interestingly, a ‘critical calcium oxalate supersaturation’ has been previously defined by Ogawa and Hatano [25] as a ΔG value of 2.8, which converts to a SS value of 8.72. Its proximity to the highest SS(COM) value of 8.40 recorded in the present study (Table 1, Fig. 1A), is noted.

Based on the results of the present survey, it would be foolhardy to claim that SS is the cause of stone formation simply because average values are significantly greater in stone formers than in healthy individuals. Given the failure of SS to convincingly discriminate between healthy individuals and stone formers, one is forced to challenge the notion of it being the critical determinant in stone formation. Supersaturation of a particular salt is certainly crucial in determining whether its precipitation will occur, and this is true for any chemical solution, irrespective of its complexity. Kavanagh [14, 26] has suggested that the limiting factor for demonstrating SS as the crucial determinant in stone formation lies in our inability to incorporate the full range of potential calcium ligands into its calculation. He cites crystallisation inhibitors and promoters in general [26] and macromolecules (glycosaminoglycans and proteins) in particular [14], in this context. Thermodynamic stability constants for calcium complexes with macromolecules are needed for these calculations, but they are not readily available. Therefore, exploration of the influence of macromolecules on the calculation of SS values for calcium salts remains untested at this time.

The present study shows that SS values of salts in urine, as they have been calculated for the past 40 years, do not yield sufficiently consistent results to make a compelling case that it is this physicochemical parameter which distinguishes stone formers from healthy individuals. Certainly, average SS values for COM, Bru and UA differ significantly between the groups (except for SS(UA), filtered; Table 2), but the mere fact that these differences exist does not necessarily mean that it is these parameters that ultimately determine stone formation or not. Indeed, evidence from several studies can be interpreted as refuting this causal argument. For example, statistically significant differences were not found between healthy individuals and stone formers for SS(COM) [21, 27], SS(Bru) [27] or SS(UA) [21, 27] or for ΔG(COM) [28]. The failure to find a significant difference between SS(COM) in recurrent and non-recurrent CaOx stone formers provides further such evidence [22]. The occurrence of supersaturation (frequently at relatively high levels) in normal individuals, without the consequence of stone formation, provides yet another argument in support of SS being a casual factor rather than a causal one in urolithiasis.

The finding in the present study that SS per se cannot distinguish between healthy individuals and stone formers does not downplay the role of urinary inhibitors. When the concentrations of these are decreased as a consequence of any particular pathological condition, SS values may increase. As critical SS values vary from one individual to another, stone formation may ensue when this occurs.

Despite its shortcomings in terms of its inability to distinguish repeatedly and consistently between healthy individuals and stone formers, SS remains the most comprehensive of all the numerous physicochemical risk factors that have been employed in urolithiasis research [29]. Unlike most, if not all other risk indices, it applies a holistic approach involving multiple urinary components in a synergistic environment and takes into account multiple chemical interactions and equilibrium processes. There can be no denying whatsoever Finlayson's assertion that ‘supersaturation is the sine qua non for stone formation’ [4]. As such, it is an essential part of the clinical evaluation of individual patients and an essential tool in urolithiasis research. However, despite it being an essential requirement for stone formation, supersaturation appears to be a casual factor rather than a causal one.

Conflicts of interest

None declared.


The author gratefully acknowledges financial support from the South African National Research Foundation, the South African Medical Research Council and the University of Cape Town. Thanks are also expressed to Dr Shameez Allie-Hamdulay for performing EQUIL2 calculations.


calcium oxalate monohydrate






uric acid