In developing world, herbal remedies are gaining popularity and 80-85% of their population relies upon alternative system of medicine. More than 7000 medicinal plants are currently in use by different cultures of the world to treat various conditions and almost 800 plants in the Asian subcontinent are shown to possess hypoglycemic activity [1, 2]. Among the various medicinal plants, bitter melon (family: Cucurbitaceae, genus: Momordica, species: Charantia) is the most popular and commonly used herb for treating diabetes [3]. There is uncertainty about the exact location of its origin, but it is cultivated in many tropical regions of India, Pakistan, China as well as South America and East Africa. Although, all parts of the plant have been shown to possess antidiabetic activities, the most commonly used part is the fruit pulp which is used in the form of juice or dried powder [4, 5]. A cluster of animal studies have documented the hypoglycemic effects of bitter melon fruit extracts [6–9]. Studies also indicate that glucose lowering effects of bitter melon fruit extract are comparable to some oral hypoglycemic agents used in the initial management of diabetes [10, 11]. Bitter melon is believed to exert its hypoglycemic effect through multiple mechanisms involving the stimulation or inhibition of the key enzymes of metabolic pathways. For example, it stimulates the key enzymes of hexose monophosphate pathway [12], increases the utilization of peripheral and skeletal muscle glucose [6, 7], prevents glucose uptake by intestine [13–15], inhibits gluconeogenesis [12, 16] and adipocytes differentiation [17] and normalizes the islet β cells [18, 19].

As compared to animal studies, only few clinical trials have been conducted so far regarding the hypoglycemic potential of bitter melon and majority of these studies lack proper control and have conflicting results [20–22]. Indeed, Diabetes United Kingdom (UK) has warned the people regarding the use of bitter melon capsules because of lack of information about its safety and recommended dose [23]. Because of the frequent use of processed bitter melon in the treatment of diabetes mellitus and the lack of information about its efficacy and safety, there is an urgent need to conduct a clinical trial comparing the hypoglycemic and/or antiatherogenic potential of bitter melon with oral hypoglycemic agent (OHA) and the present study is an attempt to evaluate and compare the antidiabetic and antiatherogenic effects as well as the safety of two different doses of bitter melon with glibenclamide in type 2 diabetic patients.

Table  1 shows the general characteristics of study groups (after cutting outliers) at baseline. There were no significant differences in the mean values of variables among the groups. Subjects in all the groups were overweight with the males higher in ratio than females. The proportion of subjects who completed the study was 94% each for group I and II and 97% for group III.

Changes in glycohemoglobin

Compared to baseline, we observed significant decreases in the mean levels of HbA1-c at the end point among patients of group I, group II and group III with the later showing greater reduction than the formers (8.25 ± 0.70 vs. 7.40 ± 0.50, p ≤ 0.05; 8.30 ± 0.55 vs. 7.15 ± 0.60, p ≤ 0.02; 8.45 ± 0.60 vs. 6.90 ± 0.75, p < 0.005 respectively). Between groups comparison at the end point revealed no significant difference (Figure  1).

Changes in plasma sialic acid

PSA showed a trend towards decreased level from baseline to endpoint among patients of group I and group II, but the temporal pattern of these reductions was different among the groups. In group I, the level decreased from 74 ± 8.10 mg/dl at baseline to 68 ± 6.70 mg/dl at the endpoint (p < 0.08), whereas in group II, statistically significant decrease occurred from 75.30 ± 8.20 mg/dl at baseline to 63 ± 6.70 mg/dl at the endpoint (p < 0.01). Moreover, the decrease was more rapid in group II compared to group I. In group III, the level slightly deteriorated and rose from 73.90 ± 7.20 mg/dl at baseline to 75.10 ± 5.20 mg/dl at the endpoint (p ≤ 0.11). The mean difference was significant at the endpoint between group II and group III (P ≤ 0.05) (Figure  2).

To the best of our knowledge, only few randomized, controlled clinical trials have been conducted to evaluate and compare the hypoglycemic effect of bitter melon fruit pulps with commonly used oral hypoglycemic agents, but most of these studies have conflicting results about the hypoglycemic effect of bitter melon and there is lack of information about its effect on diabetes associated CV risk factors [20, 26]. Results of the present study indicate significant reductions in HbA1-c levels from baseline to endpoint among patients receiving bitter melon (group I, group II) and glibenclamide (group III) with the group III showing greater reduction. At the endpoint, the difference in mean reduction of HbA1-c levels between the groups was non-significant. These results are in agreement with the recent findings by Fuangchan et al. [22] which show a significant decrease in fructosamine level of bitter melon treated type 2 diabetics but in contrast with the previous study reporting a non-significant decline in HbA1-c levels of type 2 diabetics after taking bitter melon extract for 3 months [27]. However, the study has been criticized due to lack of information about the standardization of bitter melon product. Differences in the processing of bitter melon preparation may affect the activity and the hypoglycemic effect of aqueous extract of fruit is reported to be more than the dried fruit powder [21].

A significant decrease in PSA level from baseline to endpoint was observed only in a group receiving bitter melon 4 g/day. Sialic acid is present in glycoconjugates as terminal monosaccharide and is a possible risk factor for CVD in diabetics as well as in general population [28–30]. Its high plasma level reflects the ongoing atherosclerotic process [31]. Earlier, we have shown significant improvement in PSA and cholesterol levels of type 2 diabetics following bitter melon fruit juice and the failure of dried powder of bitter melon 2 g/day (group I) to produce significant reduction in PSA confirms the hypothesis that aqueous extract of the bitter melon fruit is more efficacious than the dried fruit powder [21]. Although, we do not have a solid reason but it might be possible that drying and further processing during powder preparation may cause loss and/or inactivation of some active ingredients notably the polypeptide-p. The group receiving glibenclamide showed a slight increase in the level of PSA from baseline to endpoint. There is a direct relationship between sialic acid and blood lipids and the increase in PSA level may be the result of unfavorable effects of glibenclamide on blood lipids.

The mean changes in 2 hour plasma glucose levels after OGTT were non-significant whereas the changes in FPG levels were significant in group I and II. In group III, the mean changes in 2 hour plasma glucose after OGTT and FPG levels from baseline to endpoint were significant. The previous findings by Srivastava et al. [21] and Fuangchan et al. [22] also showed significant reductions in FPG levels in bitter melon treated subjects. Recently conducted randomized, clinical control trial report no significant changes either in OGTT or FPG levels of type 2 diabetics after taking bitter melon for 4 weeks [22]. However, the use of hot air oven for drying fruit pulp (60°C for 24 hours) may caused denaturation and inactivation of polypeptide-P and other compounds thereby decreasing the hypoglycemic effect of bitter melon [25, 32]. In addition, the lower charantin content/g of the dried fruit pulp in the study by Fuangchan et al. [22] than the present study might be due to the use of heat for drying purposes.

Unlike the group receiving glibenclamide in which the blood lipid levels deteriorated, the groups receiving bitter melon showed favorable changes in the blood lipid levels from baseline to endpoint, but did not reach the statistical significance except the triglyceride level of a group receiving bitter melon 4 g/day (group II). Moreover, the mean TC/HDL ratio was found to be low among patients of group I and group II than group III. TC/HDL ratio is more sensitive and specific index of CV risk than TC alone, specially with values >6 [33]. According to Framingham Heart Study, total cholesterol-to-HDL ratio of 5 signifies for men an average risk of heart disease whereas for women, a ratio of 4.4 signifies an average risk [34]. Treatment with bitter melon 2 g/day and 4 g/day was also associated with decreases in body weight and SBP from baseline to endpoint, but the decrease was significant only in the body weight of a group receiving bitter melon 4 g/day. These wanted effects were not observed with glibenclamide that rather exerted deteriorating effect.

Controversies about the hypoglycemic effects of BM exist due to differences in the processing of its preparation and will continue to exist until a validated standardized method for the preparation of BM formulation is not developed. The present study favors the hypoglycemic effect of bitter melon among type 2 diabetics at the dose of 2 g/day and 4 g/day however, the glycemic control was poor compared to glibenclamide. In contrast, the effect on diabetes associated CV risk factors was more favorable by BM than glibenclamide, suggesting the BM to be less hypoglycemic but more antiatherogenic than glibenclamide.