Analytical tools for determination of new oral antidiabetic drugs, glitazones, gliptins, gliflozins and glinides, in bulk materials, pharmaceuticals and biological samples

Abstract The review presents analytical methods for determination of new oral drugs for the treatment of type 2 diabetes mellitus (T2DM), focusing on peroxisome proliferator-activated receptor gamma agonists (glitazones), dipeptidyl peptidase 4 inhibitors (gliptins) and sodium/glucose co-transporter 2 inhibitors (gliflozins). Drugs derived from prandial glucose regulators, such as glinides, are considered because they are present in some new therapeutic options. The review presents analytical procedures suitable for determination of the drugs in bulk substances, such as pharmaceuticals and biological samples, including HPLC-UV, HPLC/LC-MS, TLC/HPTLC, CE/CE-MS, spectrophotometric (UV/VIS), spectrofluorimetric and electrochemical methods, taken from the literature over the past ten years (2006-2016). Some new procedures for extraction, separation and detection of the drugs, including solid phase extraction with molecularly imprinted polymers (SPE-MIP), liquid phase microextraction using porous hollow fibers (HP-LPME), HILIC chromatography, micellar mobile phases, ion mobility spectrometry (IMS) and isotopically labeled internal standards, are discussed.


Introduction
Type 2 diabetes mellitus (T2DM) is a worldwide problem affecting approximately 8% of the adult population, with predictions of more than 400 million cases by 2030 [1]. The prevalence of T2DM implies an urgent need for new treatments and preventative strategies. The disease results from progressive β cell dysfunction in the presence of chronic insulin resistance, leading to a progressive decline in plasma glucose homeostasis. Increased glucagon secretion, gluconeogenesis, renal glucose reabsorption and reduced incretin response are then observed.
Treatments recommended by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD) include drugs affecting all of the above processes [2]. In most patients, lifestyle changes and metformin (MET) from biguanides are recommended after diagnosis unless contraindications are present. If the therapeutic goal is not achieved after approximately three months, one of four oral treatment options can be considered in combination with MET: sulfonylureas (SUs), peroxisome proliferator-activated receptor gamma (PPARγ) agonists (glitazones), dipeptidyl peptidase 4 (DPP4) inhibitors (gliptins) or sodium/glucose cotransporter 2 (SGLT2) inhibitors (gliflozins). In patients with contraindications for MET, the initial drug one of these four types of drugs will be the initial treatment option. The choice of the treatment is always based on a particular patient and drug properties, with the goal of improving glycemic control and minimizing side effects [2].
The present review examines analytical methods used for the determination of giltazones, gliptins and glifozins, the second choice drug options for oral treatment, excluding MET and SUs. Additionally, detection of glinides, relatively new drugs that act as prandial glucose regulators, is discussed. Glinides are not yet included in the recommendations of ADA and EASD, but they are relatively new drugs, preferred for some patients.

Glitazones
Glitazones improve insulin sensitivity in skeletal muscles and reduce hepatic glucose production. The side effects of glitazones include weight gain, fluid retention leading to heart failure in predisposed individuals and increased risk of bone fractures. They do not increase the risk of hypoglycemia and may be more durable in their effectiveness than SUs and MET. Between them, the most frequently used is pioglitazone (PIO) (Fig. 1) while another agent is rosiglitazone (ROS). ROS was withdrawn from the market in 2010 owing to concerns of increased myocardial infarction risk. Recently, however, Food and Drug Administration (FDA) released a drug safety communication involving dispensing restrictions for ROS. This decision was based on a re-evaluation of the endpoints of Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes trial [3,4].

Gliptins
Gliptins constitute a newer class of agents for treatment of T2DM via the inhibition of DPP4, the enzyme that rapidly inactivates the incretin hormones such as glucagon-like peptide 1 (GLP1) and glucose dependent insulinotropic polypeptide (GIP). GLP1 and GIP serve as important prandial stimulators of insulin secretion and regulators of blood glucose concentrations. Therefore, inhibition of DPP4 prolongs the activity of endogenous GLP1 and GIP, decreasing the elevated blood glucose concentration in diabetic patients. The main drugs from this group are sitagliptin (SIT) (Fig. 1), vildagliptin (VIL), linagliptin (LIN) and saxagliptin (SAX) [5].

Gliflozins
Gliflozins are the newest class of drugs for T2DM. The mechanism of their action implies binding to the SGLT2 resulting in the blockade of the glucose and sodium transport cycle. In patients treated with gliflozins, urinary glucose excretion increases which results in lowering blood glucose concentrations. The main drugs from this group are dapagliflozin (DAP) (Fig. 1), canagliflozin (CAN), ipragliflozin (IPR) and empagliflozin (EMP) [6,7].

Glinides
Glinides stimulate insulin release through mechanisms similar to that of SUs but their action may be more advantageous for some patients. Repaglinide (REP) (Fig. 1), mitiglinide (MIT) (Fig. 1) and nateglinide (NAT) are carboxylic or amino acid derivatives which close the K ATP channels in β cells, targeting a low-affinity binding site on the sulfonylurea receptor 1 subunit. Because the effects of these drugs are rapid and short-lived, they are used to curtail postprandial excursions in glucose, thus exposing patients to much less risk of hypoglycemia than SUs [8].

Analytical tools for determination of oral antidiabetics
Pharmaceutical analysis has become one of the most important stages in the therapeutic process. Drug analysis includes analytical investigations of bulk drug materials, the intermediate products, drug formulations, impurities and degradation products.
Analytical techniques play a significant role in understanding the chemical stability of the drug, in evaluating the toxicity of some impurities and in assessing the content of drug in formulations. Also, they are fundamental tools in pharmacokinetic studies where the analysis of a drug and its metabolites in biological fluids must be performed. Polypharmacy, which has become an integral part of therapy for many diabetic patients, further supports the importance of drug analyses. To support polypharmacy, methods suitable for two or more components are needed for quality control of such combined formulations as well as for assays in biological samples. This paper presents analytical procedures elaborated for the listed drugs: HPLC/LC-MS, TLC/HPTLC, CE/CE-MS, spectrophotometric (UV/VIS), spectrofluorimetric and electrochemical methods. It is based on a review of the literature from the past ten years (2006-2016).

Determination of glitazones in bulk substance and pharmaceuticals
The methods described for PIO [9,[10][11][12][13] and ROS [14] are suitable for their determination in bulk drug materials, formulations and in the presence of impurities or degradation products (Table 1). PIO is described in Ph. Eur.
[9] and a HPLC method is proposed for its determination. The methods were mostly performed using C18 columns, isocratic elution and UV detection, which were sufficiently effective [9][10][11]14]. Two methods [12,13] were based on gradient elution which allowed for effective separation of PIO and its impurities. In the study of Ramulu et al. [13] impurities were additionally identified by LC-MS/MS, 1 H NMR, 13 C NMR and IR methods.
The literature presents several HPLC methods effective for two component combinations of PIO or ROS with MET [15][16][17], SU [18][19][20], atorvastatin [21] and olmesartan [22]. Also, three component formulations with PIO or ROS, MET and SU or some cardiovascular drugs were analyzed and quantified [23] ( Table 2). In the pharmaceutical industry, a significant problem is the validation of cleaning procedures in the manufacturing arena. To detect contamination from manufacturing equipment, a fast isocratic HPLC method was developed to separate and detect six antidiabetic drugs --PIO, . The HPLC method of El-Refay et al. [15] was based on micellar mobile phase containing sodium dodecyl sulfate (SDS) at concentration above critical micellar concentration. In this way simultaneous determination of hydrophobic and hydrophilic analytes like PIO and MET could be achieved easier than using conventional HPLC method.
A HPLC method with MS/MS detection and positive electrospray ionization (ESI) was elaborated by Wang et al. [23] to determine ROS, phenformin and 2 SUs as illegal additives in dietary supplements and traditional medicines.

Determination of gliptins in bulk substance and formulations
A few HPLC methods were described for determination of gliptins, including SIT [9,25,26], SAX [27] and VIL [28,29], in bulk substances and single component formulations ( Table 1). The cited methods were based on C18 or C8 columns, isocratic elution and UV detection. It is worth mentioning that the method of El-Bagary et al. [25] was based on fluorescence (SF) detection, which was ca. 20 times more sensitive than the alternative LC-UV procedure.
Two methods were suitable for determination of gliptins in the presence of impurities [9] or synthetic intermediate products [28]. Identification of the related  (Table 3). Some reports showed the suitability of CN [9,37] and C8 columns [32,34], because of relatively high polarity of these drugs. At the same time, ultra performance liquid chromatography (UPLC) was successfully applied for the assay of SIT by Malleswararao et al. [34] (Fig 2).
Some of the presented methods included stress degradation studies of the drugs. The proposed procedures occurred suitable for determination of the analytes in the presence of their degradation products, so they could be described as the stability-indicating methods [25,31,32,35,37,38]. The study of Narendra and Jeyabalan [35] resulted in detection of 15 degradation products of SIT.

Determination of gliflozins in bulk substance and formulations
Gliflozins are the newest oral antidiabetics , having been introduced as therapeutic interventions for T2DM only in 2015. Therefore, in the literature only several methods have been described for their determination in bulk substances [39,40] (Table 1) and combined formulations with MET [41,42] (Table 3). All methods were based on C18 columns, isocratic elution and UV detection. At the same time, the stability-indicating property of the last method was proven.

Determination of glinides in bulk substance and formulations
REP and NAT are present in official monographs [9], and respective HPLC methods are proposed for their analysis ( Table 4). Determinations of REP and NAT in bulk drug materials and formulations were mainly performed using C18 columns, isocratic elution and UV detection [43][44][45][46][47][48][49].
In combined formulations, REP and NAT are mainly mixed with MET and SUs. Suitable analytical tools for their determinations were reported by several authors [45][46][47][48].
The HPLC method of El-Wasseef [48] was based on the micellar mobile phase containing SDS. The involvement of the micellar mobile phase made simultaneous determination of hydrophobic compounds such as NAT and SU in the presence of hydrophilic substance like MET more reliable (Fig. 3).

CE methods for determination of oral antidiabetics in pharmaceuticals
Although CE is based on a different mechanism of separation than HPLC, it is sufficiently precise, reliable and robust for quantitative measurements. As far as analysis of glitazones, gliptins, gliflozins and glinides is concerned, CE-based methods were used for assays in pharmaceuticals, biological samples and for enantioselective separation.
In the paper of Piccoli et al. [50] successful separation of VIL and MET was achieved by CE method with UV detection at 207 nm, using a positive polarity of 25 kV, hydrodynamic injection and the running buffer consisted of sodium tetraborate. The method was linear in the range of 30-60 μg/ml with precision lower than 3%. Additionally, specificity and stability-indicating capability of the method was proven by enforced degradation studies.

Enantioselective separation of oral antidiabetics
If one of the enantiomers of a chiral drug has a significantly better effect or fewer side effects than another, development of an enantiomerically pure preparation is an important goal. However, in order to perform the necessary investigations, the pure enantiomers must either be synthesized or separated from the racemic mixture. Chromatography, especially HPLC, has become an essential tool for such chiral separation. However, some HPLC methods are time-consuming and are not suitable for fast analysis in control laboratories. Therefore, CE techniques were examined and performed similarly to HPLC or were even found to be more effective at separating enantiomers. The literature holds examples of the use of both methods for resolving racemic mixtures for several antidiabetic agents, including PIO, ROS, SIT, VIL, REP and NAT.

HPLC methods for chiral separation of antidiabetics
The individual enantiomers and racemates of glitazones have shown equivalent activity as antidiabetic agents; therefore, both the drugs are present on the market as racemic mixtures. HPLC methods for chiral separation  of (R)-and (S)-enantiomers of PIO and ROS have been developed using a chiral stationary phase based on cellulose and amylose derivatives [51-53] ( Table 5). In addition, Calixto and Bonato [53] successfully separated the enantiomers of the main metabolites of ROS (N-desmethyl-ROS and OH-ROS) using a chiral stationary phase based on macrocyclic antibiotics (Fig. 4). UV detection was usually sensitive enough for such determinations [51,53]. However, one report was based on MS detection with atmospheric pressure chemical ionization (APCI) [52].
The pharmacological activity of SIT is assigned specifically to the (R)-enantiomer while the (S)-enantiomer is an impurity [9]. On the contrary, only (S)-enantiomer of VIL is in clinical use while the (R)-enantiomer is considered as an optical impurity. For selective determination of these analytes two HPLC-UV methods have been described [9,54] (Table 5). Another method involved pre-column derivatization of SIT with o-phthalaldehyde (OPA) and N-acetyl-L-cysteine to form diastereomeric derivatives which were separated on a C18 column and detected using SF detection with excitation at 330 nm and emission at 450 nm [55].
Glinides also have stereochemical considerations. REP is manufactured as a single (S)-isomer; however, the (R)-enantiomer could be present as a chiral impurity. As far as NAT is concerned, only the trans-isomer, or D-NAT, is used in therapy. The pharmacopoeial method for estimation of the enantiomeric purity of REP requires ca. 1 h for saturation of an α 1 -acid glycoprotein column [9]. Another method proposed in the literature was based on a stationary phase with an amylose derivative and was found to give faster resolution [56]. For determination of the chiral impurity of NAT, an HPLC method based on a column with urea was proposed while two other chiral impurities could be separated on a column with metacrylic gel [9]. Other method based on amylose and cellulose derivatives as chiral stationary phases [57] was successfully applied to estimate in vitro transport of NAT enantiomers in different regions of rat small intestine (Table 5).

CE methods for chiral separation of antidiabetics
CE systems were used for successful separation of the mentioned enantiomers of oral antidiabetics. The running buffers were based on sulfobutylether-α-cyclodextrin (SBEα-CD), sulfobutylether-β-cyclodextrin (SBE-β-CD) and dimethyl-β-cyclodextrin (DM-β-CD) [51,58-60] ( Table 6). When the results obtained by the CE method were compared to those obtained with respective chiral HPLC assay [51], the CE method seems to be equally selective and sensitive, however to some extent, less precise than HPLC (Table 5). Where: APCI is atmospheric pressure chemical ionization.

TLC/HPTLC methods for analysis of oral antidiabetics
TLC and HPTLC methods have been successfully applied to the separation of compounds closely related in chemical structure or to the quantification of them in simple or combined pharmaceuticals. The results obtained by researchers clearly show that these methods could be used as reliable tools for quantitative determination of oral antidiabetics as good alternatives to HPLC methods.

Chromatography in normal phase systems
Normal-phase TLC/HPTLC-UV densitometric methods were described as sufficiently effective tools for routine analysis of respective formulations containing one active component as well as for the combined ones [10,18,21,61-65] (

Chromatography in reversed phase systems
In the literature concerning oral antidiabetics, only two reports using reversed-phase chromatography were found. One procedure was described as a stability-indicating method for the determination of PIO in bulk substance and formulations (Table 7) [74]. The main goal of the second study was to obtain experimental lipophilicity data useful for prediction structure-activity relationships for five glitazones, including PIO and ROS [75]. The authors used C18 TLC plates and binary mobile phases containing water and organic modifiers, acetone (50-85%), 1,4-dioxane (40-80%) and methanol (55-95%). After development, the Where: DM-β-CD is dimethyl-β-cyclodextrin; SBE-α-CD is sulfobutylether-α-cyclodextrin; SBE-β-CD is sulfobutylether-β-cyclodextrin.

UV methods for detection of oral antidiabetics
Direct assays in the UV range (Table 8) or techniques utilizing various derivatives (Table 9) were applied for the drugs mentioned in the present review. The methods were mainly proposed as simple and inexpensive alternatives to HPLC for quantitative measurements of active ingredients in pharmaceuticals. Some of them were proven to be sufficiently selective to detect the drugs in combined dosage forms or in the presence of degradation products.  [77] was based on formation of complexes of ROS with Cu(II) and Al(III) ions, in borate buffer of pH 6.5. The the formed complexes were measured by absorbance at 318 nm with linearity in the ranges of 8-80 and 5-70 μg/ml, for Cu 2+ and Al 3+ ions, respectively.

Direct UV spectrophotometric methods
As far as other antidiabetics are concerned, a simple zero-order spectrophotometry was proposed for determination of SIT [82] and SAX [83]. For two component formulations of gliptins with MET, SU, PIO and simvastatin, the simultaneous equation methods were used [84][85][86].
In the literature, one spectrophotometric method was proposed for determination of DAP from gliflozins. The drug was determined with the sensitivity ca. 5 times higher than that obtained in the alternative HPLC method [39]. REP and NAT from glinides were determined in one component formulations by direct spectrophotometry and difference spectroscopy [87][88][89]

Derivative UV spectrophotometric methods
For determination of PIO and ROS in bulk materials and formulations, the first-order and the second-order derivatives as well as the ratio absorbance methods have been utilized. These methods have been found to be sufficiently accurate to detect the drugs alone [92] and sufficiently selective for the combined dosage forms with MET (93) or SU [94].
Similar procedures have been used to analyze SIT and LIN [36,95,96] as well as REP and NAT [87,91,97]. The method proposed by Al-Khalidi et al. [97] was successfully used for the dissolution testing of REP and was found to be sufficiently reliable, simple and fast. Additionally, a statistical comparison of the results obtained by spectrophotometry and the USP HPLC method showed that there was no significance difference between these two methods.
In the study of Lotfy et al. [98], three methods based on manipulating ratio spectra (i.e. ratio difference, ratio subtraction and induced amplitude modulation) were proposed for detection of SIT in the presence of MET. The authors also recommended three methods based on derivative spectroscopy: amplitude factor, amplitude subtraction and modified amplitude subtraction [98].

Spectrofluorimetric methods for determination of oral antidiabetics
In a study by Telny et al. [99], native fluorescence of ROS was measured in methanol at 311 nm with excitation at 297 nm. The response was found to be linear in the range of 1-5 ng/ml. Caglar et al. [100] proposed two methods based on the natural fluorescence of SIT. In the first method, fluorescence was measured at 353 nm after excitation at 259 nm. The second method was based on the reaction between SIT and fluorescamine in borate buffer at pH 9.0, and fluorescence of the product was measured at 475 nm after excitation at 390 nm. Calibration curves were constructed in the ranges of 0.5-10 μg/ml and 0.2-1.4 μg/ml for the first and the second method, respectively [100]. A specrofluorimetric method for SIT determination with emission at 575 nm and excitation at 263 nm was proposed by El-Bagary et al. [95]. The method was linear in the range of 0.25-110 μg/ml. In the study of Salim et al. [101] the native fluorescence of SIT was enhanced by phosphate buffer of pH 4.0 in the presence of SDS. Fluorescence was measured at 300 nm after excitation at 270 nm. A proposed method described by Salim et al. showed an LOD equal to 5ng/ml and was used for stability studies after exposure of the drug to stress degradation [101].
A spectrofluorimetric method was also used for estimation of REP in tablets as well as in receptor fluid during in vitro permeation studies. REP was observed to exhibit fluorescence at 379 nm after excitation at 282 nm.

VIS methods for determination of oral antidiabetics
Over the past 10 years, visible spectrophotometric (VIS) methods have been used for determination of PIO and ROS in bulk materials and respective formulations [103,104]. Generally, the procedures were based on ion pair formation or charge transfer reactions. VIS methods for the detection of gliptins can take advantage of the reaction of the primary amino group of SIT and SAX with acetylacetone and formaldehyde, which produces a yellow Hantzsch product (Fig. 6) [105]. Other colorimetric determinations of SAX and VIL have been performed using different charge transfer reactions [106,107]. VIS assays of REP have been based on formation of ion-pair complexes or the charge transfer reactions (Fig. 7) [108,109]. A simple procedure was proposed for estimation of NAT based on oxidation of the drug with 2,4dinitrophenylhydrazine [110].
Despite some limitations of colorimetric methods, the results presented in the cited papers showed that they are sufficiently accurate and precise for the analysis of drugs in pharmaceutical formulations, especially when simple and economic procedures were needed (Table 10).

Electrochemical determinations of oral antidiabetics
A few electrochemical methods for the analysis of PIO and ROS have been published.  used membrane selective electrodes for the determination of PIO in pharmaceuticals. Saber and Shah [112] developed new membrane sensors that were employed in the detection of PIO in urine, using iodobismuthite-PIO ion-pair incorporated in a PVC membrane with o-nitrophenyl octyl ether or dioctyl phthalate as plasticizers. The electrodes showed good selectivity with respect to some inorganic cations, sugars, cellulose derivatives, magnesium stearate and ascorbic acid. Cyclic voltammetry and differential pulse voltammetry methods were also proposed for PIO and ROS, using carbon paste and glassy carbon electrodes as sensors [113]. Wang and Song [114] employed a flow through voltammetry method using a gold electrode. In the study of , adsorptive cathodic stripping voltammetry was applied to the detection of PIO in pharmaceuticals and biological fluids. An LOD of ca. 3 ng/ml was achieved. Wang and Song [116] applied electrochemical impedance spectroscopy (EIS) using a silver electrode. EIS could be an attractive analytical alternative because of its high sensitivity and good characteristics, including a rapid response, a satisfactory range of linearity (0.2-40 µg/ml) and good stability.
Some authors proposed various types of solid phase extraction (SPE) [143][144][145][146][147][148]. In the studies of Jafari et al. [144] and Nageswara Rao et al. [145], molecularly imprinted polymers (MIPs) were used as packing materials in molecularly imprinted SPE (MIP-SPE). This technique is relatively a new concept for the cleaning of biological samples. MIPs possess some specific cavities previously designed the target molecules. Additionally, they are stable in low and high pH values, in a variety of solvents and over a broad temperature range, which provides for flexibility in the development of bioanalytical methods (Tables 11-12).
Determination of antidiabetics could be an analytical challenge when the simultaneous extraction of two or more compounds is needed. Extraction is especially challenging when the simultaneous measurement of glitazones and their metabolites or glitazones in the presence of MET is concerned, because of large differences in polarity. An interesting procedure that improved recoveries from such mixtures was liquid phase microextraction (LPME) using porous hollow fiber (HF-LPME) This procedure has several advantages over other extraction methods. It is very simple, rather inexpensive, provides excellent clean-up and presents low consumption of organic solvents. After optimization of the procedure a significant enrichment factor could be obtained [119, [149][150][151] (Tables 11,13-14). Where HF-LPME is hollow-fiber liquid phase microextraction; IMS is ion mobility spectrometry; MIP is molecularly imprinted polymer; SF is fluorescence detection; TBME is tert-butyl methyl ether.
In the study of Swales et al. [157], the use of laser diode thermal desorption (LDTD) MS/MS was proposed for simultaneous quantification of SIT and MET in mouse and human dried blood spots. LDTD uses an infrared laser to thermally desorb analytes coated onto the metallic surface of specially designed plates. Analytes subsequently can be efficiently ionized with an APCI interface.
An interesting method based on ESI ion mobility spectrometry (ESI-IMS) in positive mode was used for analysis of PIO in cow plasma by Jafar et al. [144]. IMS relies on the mobility of the analyte ions in the gas phase operating at atmospheric pressure. According to the authors, it is much easier to use and cost effective compared to MS techniques [144].
Although LC-MS and LC-UV detections were commonly used, other methods for detection of oral antidiabetics have also been published. Using the native fluorescence of the drugs the detection of peaks using SF detectors was proposed for ROS [132,159], SIT [25] and REP [125]. The last two methods allowed for detection at concentrations of 0.25 and 1.0 ng/ml, respectively, meaning that the fluorescence method was at least 10 times more sensitive than other methods reported.
of 0.04 ng/ml. Similar sensitivity for REP was obtained by a LC method with coulometric detection [139]. These last two reports showed the highest sensitivity among all LC methods presented for REP (Fig. 9). As was stated above, the examined drugs differ in their polarity, especially when non polar glitazones were assayed in the presence of highly polar MET. This problem was effectively overcome using special types of columns, such as HILIC [120]. This system provided high separation efficiency with good peak shape compared to reversed phase chromatography. HILIC columns were also used for determination of relatively polar compounds, such as SIT Where: CL is chemiluminescence detection; HF-LPME is hollow-fiber liquid phase microextraction; SF is fluorescence detection; TBME is tertbutyl methyl ether.
[145,147]. In the study by Rao et al. [145], zwitterionic HILIC chromatography was described. According to the authors, zwitterionic materials (e.g. polymeric sulfoalkylbetaine) were characterized by carrying both positive and negative charges on the surface. The electrostatic interactions between oppositely charged groups in close proximity weakened the interactions of the stationary phase with the charged analytes improving separation and peak shape.
Another interesting topic in biomedical analysis is the use of isotope-labeled, mostly deuterated, compounds. Such stable isotopes are ideal internal standards. They have the same solubility, extraction and chromatographic behavior as their non-labeled counterparts. On the other hand, their different molecular weights make them distinguishable from the non-labeled compounds in LC-MS. In the papers cited above such standards were used for determination of PIO and its metabolites, i.e. PIO-d4, MIV-d5, SIT-d4, SAX and OH-SAX-13 Cd-2 and REP-d5 [122,131,143,153].

CE methods for biological assays
For determination of ROS in biological fluids, HF-LPME with dihexyl ether coupled with CE method was proposed by Al Azzam et al. [119]. Fused silica capillary (63 cm x 75 µm), hydrodynamic injection at 50 mbar for 15s, 25kV voltage, capillary temperature 25°C, 25 mM sodium acetate of pH 4.0 and UV detection at 200 nm were used. The LOD thus obtained was 2.83ng/ml while the precision was estimated as 8.44%. The authors showed that the CE procedure was equally precise as HPLC, but, to some extent, less sensitive. For simultaneous determination of ROS and MET in human plasma, deproteinization and CE coupled with MS/MS detection was developed by Znaleziona et al. [128]. This method combined the advantages of both techniques, leading to high efficiency, sensitivity and selectivity, and it allowed effective separation of the drugs within 11 min (Fig. 10). The authors used 50 mM formic acid as a running buffer, positive 20kV polarity and positive ESI with selective ion monitoring (SIM) mode. The sheath liquid was composed of water, methanol and formic acid. The LOD value obtained by the authors, i.e. 4.42ng/ml, was sufficient for therapeutic monitoring of the drug [128].

Other methods for biological assays
In the study of Adhikari et al. [130], direct UV spectrophotometry was used for determination of ROS in rat and human plasma after simple deproteinization of the samples with acetonitrile. However, the sensitivity of this method equal 0.72μg/ml seems to be insufficient in comparison with HPLC methods proposed in the literature (Table 11 and Table 15).
The fluorimetric method proposed by Walash et al. [77] for ROS was based on the complex formation with Al 3+ ions in acetate buffer at pH 5.0. The fluorescence measured at 376 nm after excitation at 318 nm, was double the native fluorescence of the drug. At the same time the method was ca. 100 times more sensitive as simply measuring the absorbance of the complex in borate buffer at pH 6.5. The linearity of the method was obtained in the range of 0.03-2.0 μg/ml. The method was applied for determination of ROS in spiked and real human plasma samples.

Conclusions
In the literature, a broad range of methods have been presented for the analysis of oral antidiabetics in bulk materials and pharmaceuticals. HPLC with UV detection and UV spectrophotometry were mainly used, due to their accuracy, precision and sensitivity. These methods were adequate to analyze the drugs in single component formulations as well as in combinations. Also, TLC/HPTLC with densitometric detection and VIS spectrophotometry were widely used for typical analysis in pharmaceutical formulations. The former method was frequently proposed as an alternative to HPLC while the latter method was used when simplicity and cost effectiveness were required.
HPLC-UV and LC-MS are undoubtedly the methods of choice for bioanalytical assays. Bearing in mind the data presented above it could be concluded that higher selectivity of LC-MS was the main difference between the described methods, while sensitivity was similar. Additionally, the LC-MS systems were frequently realized as UPLC techniques. All authors confirmed that UPLC drastically reduced the mobile phase consumption, thus having obvious economic and ecological consequences. At the same time, significant improvements in resolution and sensitivity were achieved.
Other methods such CE were also used but they were designated for special analytical purposes, e.g. separation and determination of enantiomers. In turn, electrochemical methods were adequate and helpful when separation and extraction of the active substances from the matrix were not needed.
To our knowledge, this paper is the first complete review concerning the analysis of new oral antidiabetics, including glitazones, gliptins, glifozins and glinides. A few available reports describe the analyses of individual drugs, such as PIO [160] and NAT [161]. Also, a review concerning the analysis of gliptins exists, but it was based on the papers published over a briefer time frame (2012-2014) [162].
Undoubtedly, glitazones, gliptins and glinides are widely examined with analytical techniques, in contrast to gliflozins which are the newest therapeutic group for T2DM. On the other hand, all groups of antidiabetics are still extensively developed, and almost every year a new drug similar to these described here appears on the market. The analytical methods, gathered together and compared, can facilitate steps in designing, in examining, in manufacturing and in controlling already discovered drugs as well as new substances.