Oral dryness and reduced salivary flow rates are some of the most predominant and troubling oral sequelae of pSS. This discussion focuses on the hypothesis if impaired salivary flow and/or composition and high caries experience may serve as potential markers of disease severity in pSS.
Caries experience, oral hygiene, and salivary flow rates
The present study confirms the results of previous cross-sectional studies showing that pSS patients have a significantly higher DMFS compared to healthy controls, despite the fact that these patients often have a good oral hygiene and frequent dental follow-up visits [5, 9, 10]. When comparing SS patients with patients suffering from other immune diseases and patients with xerostomia of other origins, Boutsi et al.  found no significant differences in the number of decayed, missed or filled teeth. In accordance with our results, the SS patients were characterised by having lower salivary flow rates, better oral hygiene habits, slightly higher gingival scores, but similar plaque scores, compared to the other groups . Regarding the other periodontal measures, our results support those of previous studies showing that presence of periodontal disease is not substantially increased in pSS [2, 26]. The slightly increased gingival scores could be explained by altered inflammatory response probably due to hormonal changes . It has been reported that pSS patients harbour lower numbers of periopathogenic microorganisms than healthy controls . In contrast, and despite a good oral hygiene, pSS patients appear to harbour higher numbers of cariogenic and acidophilic microorganisms such as Streptococcus mutans and Lactobacillus species than healthy controls . Similar relationships between low salivary flow rates and increased Lactobacillus counts have also been observed in patients with hyposalivation of other origins . Thus, impaired salivary flow and changes in saliva composition as seen in pSS are assumed to favour a more aciduric oral microflora manifested by an increased incidence of caries and fungal infections . At low flow rates, the bicarbonate concentration, pH, and buffer capacity as well as the clearance of microorganisms and dietary sugars in the oral cavity generally decrease , thereby promoting an environment dominated by these oral pathogens and prolonged exposure of dietary sugars to the teeth. In pSS patients with severely reduced salivary flow rates the shift in the oral microflora appears to occur despite a good oral hygiene. In addition, the high number of microbial retention sites generated by dental restorations such as fillings, crowns and bridges found in pSS patients may contribute to the shift in oral microflora  and complicate the maintenance of sufficient oral hygiene procedures.
In accordance with previous studies we found that the DMFS score is inversely correlated to salivary flow rates and especially the unstimulated whole saliva flow rate [8, 10]. However, it has not yet been possible to identify a cut-off value for salivary flow rate that can predict the risk of developing dental caries, as caries is a multifactorial disease. In this study, 90% of the patients had UWS flow rates below 0.10 ml/min, but as this value is merely a part of the classification criteria for pSS, it does not necessarily reflect the point at which caries is likely to develop. It has been suggested that salivary flow rates up to 0.16 ml/min, which include most pSS patients, may result in oral candidiasis  and increased development of experimental caries . Meanwhile, the onset of increased caries activity in pSS remains unclear. It has been stated that in patients with SS, loss of teeth due to caries precede the first symptom of xerostomia by on average 9 years . Changes in salivary flow and composition, and subsequently development of caries, appear to precede the symptom of oral dryness by several years. Increased caries activity that cannot be explained by changes in habits related to oral hygiene or diet may therefore represent a useful clinical feature to suspect early pSS in women without complaints of oral dryness, or intake of xerogenic medications. In contrast to the uncertainty pertaining to the onset of pSS and the initial consequences in terms of increased caries lesions, the future perspectives on the patients' dental health seem gloomy without professional preventive intervention. In accordance with the observation that past caries experience is one of the best predictors for future caries [32, 33], the pSS patients in this study have a much higher risk of developing future caries and ultimately to loose teeth than the control group. Accordingly, these patients should receive an individual dental care programme in terms of oral hygiene instructions, professional oral hygiene regimens, fluoride treatment, dietary supervision and frequent dental follow-up visits in order to prevent accelerated caries development.
Self-assessment of oral dryness
In healthy subjects, the sensation of oral dryness usually occurs when whole saliva flow rate is reduced with more than 50% . In this study, all patients had complaints of dry mouth and also substantially decreased salivary flow rates, whereas the healthy controls had no dry mouth complaints and in general, salivary flow rates within the normal range .
Nonetheless, in pSS, the scores of oral dryness were only inversely correlated to SWS, which may reflect a larger span of SWS values than of UWS. We have previously found an inverse correlation between scores of oral dryness and UWS . In the present study, the patients had symptoms of oral dryness median 10 years prior to the diagnosis of pSS. Symptoms of oral dryness in combination with the participants' age could explain about half of the variance in DMFS (Table 5). The questionnaire may therefore not only be helpful in assessing the intensity of oral dryness but also in identifying patients with high DMFS.
Saliva pH and buffer capacity
The ability of human saliva to buffer acids is essential for maintaining pH values in the oral environment above the critical pH for hydroxyapatite (HAP), thereby protecting the teeth against demineralisation. The buffer systems responsible for the human saliva buffer capacity include the bicarbonate, phosphate and protein systems [17, 36–38]. In normalcy, where the pH ranges from 6.0 to 7.5, the bicarbonate and phosphate buffer systems are by far the dominant ones having optimal buffering capacity at their pK values of 6.1 and 6.8, respectively [16, 19], whereas the proteins have some effect on the buffer capacity at acidic pH values below 5 [16, 38].
In this study, the parotid saliva buffer capacity was calculated individually for each pSS patient and healthy control based on the saliva bicarbonate and phosphate concentrations at the respective saliva pH values. This calculation gives an estimate of the buffer capacity of the saliva at the time it is secreted from the parotid gland. In the pSS patients, the buffer capacity of UPS was significantly lower than in the healthy controls, mainly due to the low pH value and bicarbonate concentration in saliva caused by the low flow rates. Moreover, in pSS most of the pH values of UPS were below the relevant pK values of both the bicarbonate and phosphate buffer systems. As compared to healthy controls, the pSS patients will therefore experience far more abundant pH drops in their saliva if exposed to acidic challenges leading to a higher risk of tooth demineralisation. Apart from a low buffer capacity, and thereby impaired ability to maintain a non-acidic saliva pH, the pSS patients also had a significantly lower degree of saturation with respect to HAP in their saliva (Table 4). Thus, in pSS, the mean pH of UPS was only one tenth of a pH unit above their mean critical pH. This implies that even a minor drop in pH will lead to undersaturation of their saliva with respect to HAP and result in either caries lesions or erosive damage to the teeth depending on the origin of the acidic challenge.
Salivary bicarbonate and phosphate
Despite the differences between pSS patients and healthy controls in sodium and chloride concentrations at comparable low flow rates, stimulated or not, the HCO3
- concentration and saliva pH did not differ between the two groups (Fig. 1A–D). The transport of bicarbonate in the salivary glands is believed to occur via chloride/bicarbonate exchange mechanisms . The concentration of bicarbonate in saliva is a consequence of the metabolic CO2 -turnover in the salivary glands. CO2 freely diffuses across the epithelial boundaries, and due to the presence of carbonic anhydrase, the partial pressure for CO2 and pH in the glandular compartments governs how much of bicarbonate buffer system is present in form of HCO3
- in the saliva. The duct epithelium has dual functions with respect to bicarbonate transport, since it can both reabsorb (at low secretion rates) and secrete (at high secretion rates due to increased metabolic turnover of the gland).
This study demonstrated a tendency towards lower total phosphate concentrations in pSS patients than in healthy controls. Previous studies have found significantly reduced phosphate concentrations in stimulated parotid and SM/SL saliva of patients with SS compared to healthy controls and patients with conditions resembling SS [40–42]. It should be stressed that the mechanism behind phosphate transport in human salivary gland tissues has not yet been fully characterised. It probably includes an acinar secretion and/or a ductal reabsorption via sodium-phosphate co-transport mechanism as that observed in the renal proximal tubules .
Salivary sodium and chloride
Our sialochemical results are interpreted within the frame of the classical two-stage model of saliva formation . Under normal physiological conditions, and in response to nervous stimuli, the acinar cells produce primary saliva, which has an ionic composition resembling that of plasma. As the primary saliva passes through the duct system it becomes modified by reabsorption of sodium and chloride (but without water due to the low water permeability) whereby the final saliva secreted into the oral cavity becomes hypotonic with sodium and chloride concentrations much below that of the original primary saliva. The composition of the final saliva secreted into the oral cavity strongly depends on the secretion rate in such a way that at low flow rates the saliva contains low sodium and chloride concentrations and as the flow rates increase the concentrations of sodium and chloride will rise. This normal physiological relation between parotid flow rate and sodium and chloride concentrations was seen in the healthy controls and in some of the pSS patients (0.01–1.00 ml/min) as well (Fig. 1A and 1B). However, for other pSS patients with flow rates within the same frame, another picture emerged, since their sodium and chloride concentrations were remarkably higher than normally whether stimulated or not. Thus, on a group basis the pSS patients have significantly higher concentrations of sodium and chloride than the healthy controls (Table 4). This finding is in accordance with several previous studies on whole saliva, parotid and submandibular/sublingual saliva (SM/SL) [5, 40–42, 45, 46]. The concentrations of sodium and chloride have also been shown to be higher in SM/SL of patients with pSS and secondary SS compared to patients with clinical conditions resembling SS, i.e., sialoadenosis, sodium retention dysfunction syndrome and medication-induced xerostomia . Overall, these compositional changes appear to be unique for some pSS patients. It has been stated that SM/SL glands are affected earlier by SS than the parotid glands due to an average reduction of stimulated SM/SL flow rate preceding that of the stimulated parotid flow rate [40, 42]. On the other hand, cut-off values for sodium, chloride and phosphate in SPS and stimulated SM/SL being predictive for SS demonstrated almost similar specificity (69 and 71%, respectively) and sensitivity (81%) .
Other salivary constituents
Despite the low salivary flow rates seen in the pSS, the acinar transport mechanisms involved in the formation of primary saliva seem to be unaffected by the glandular lymphocytic infiltration. Accordingly, concentrations of potassium, total calcium, total protein and amylase activity in whole and parotid saliva did not differ from those of the healthy controls, which is in agreement with previous reports [40, 45]. Furthermore, it has been shown that the output of statherin and acidic proline-rich proteins, which reflect the secretion of selected parotid proteins, did not differ between pSS patients and healthy controls . Normal concentrations of total calcium, total protein and levels of amylase activity indicate that the remaining functional acinar cells are capable of synthesis and secretion of primary saliva with normal composition despite the marked lymphocytic infiltration and structural changes.
Impaired in ductal salt reabsorption reflects disease severity in pSS
The changes in salivary composition indicate that the duct epithelium, and mainly the striated duct epithelium, cannot effectively reabsorb the high concentrations of sodium and chloride of the primary saliva in some of the pSS patients, despite low salivary flow rates. The pSS patients who exhibited high concentrations of sodium and chloride were also characterised by having the lowest flow rates, the highest focus score (FS) and highest concentrations of serum autoantibodies (AB) in addition to a tendency of being younger than the pSS patients with normal salivary concentrations of sodium and chloride. Overall, the subgroup of pSS patients with the highest concentrations of salivary sodium and chloride and lowest total phosphate concentrations appeared to be more severely affected by pSS having more exocrine and non-exocrine disease manifestations than those without these salivary changes supporting previous observations . It should be stressed that there was no relationship between reduced salivary flow and compositional changes and intake of medications. The question is whether the subgroup of patients with normal salivary concentrations of sodium and chloride and some preserved salivary gland function has a late onset of pSS or a "milder" glandular response to autoimmunity compared to the subgroup with high salivary sodium and chloride concentrations or salivary secretion close to or equal to 0 ml/min. Another relevant question is whether the latter subgroup of patients also has a greater risk of developing malignant lymphoma. The results of our study therefore need to be tested in a large prospective cohort study and compared with group of patients with non-immunological destruction of their salivary glands in order to validate the use of specific salivary changes and dental caries as markers of salivary gland dysfunctional severity/disease severity in pSS. At present there is no international expert consensus regarding measures for assessment of disease activity, severity, damage or outcome in pSS that can be used in the evaluation of clinical trials of new therapies and longitudinal observational studies. Recent reports, however, indicate that the development and evaluation of such measures has begun [47, 48].
In pSS, the salivary gland histopathology is characterised by periductal lymphocytic infiltration and acinar destruction. The duct epithelium, however, appears relatively unaffected by the lymphocytic infiltration, which contrasts the observed salivary changes in pSS. On the other hand, the discrepancy may be explained by morphological differences between the labial salivary glands, which predominantly consist of mucous acini and have a very short duct system, and the parotid glands, which comprise serous acini and have long duct system. Nevertheless, it has been stated that the histopathological changes of the labial salivary glands mimic those of the parotid glands [49, 50]. It has been suggested that Bcl-2 positive basal cells of striated/excretory ducts possess an extensive capacity for pluridirectional morphogenetic differentiation . On this basis, it could be speculated that the duct cells in some pSS patients possess antigenic properties, which initiate an autoimmune response that could be associated with morphogenesis and cell differentiation of the salivary gland tissues. It has not yet been possible to identify a specific anti-salivary duct antibody in pSS that is capable of inhibiting cell differentiation. Our finding of more pronounced acinar and ductal tissue functional impairment in pSS patients with both FS and AB, than in patients with FS or AB, supports the idea that circulating autoantibodies or inflammatory mediators produced locally by the inflammatory cells, interfere with the neural release of neurotransmitter substances or interact with the binding of neurotransmitters to receptors on the cell surface, thereby impairing the acinar secretion and/or the ductal reabsorptive modification of saliva . Along this line, an in vitro study on isolated human acini and duct segments from pSS patients have shown that these cells possess functional receptor systems and normal response in changes in the intracellular free calcium concentration upon maximal secretagogue stimulation . The fact that the ductal uptake of sodium via amiloride-sensitive epithelial sodium channel (ENaC) is regulated by circulating adrenal mineralocorticoids, e.g. aldosterone  could also indicate that some pSS patients have low levels of aldosterone. However, this still needs to be clarified.