Penelitian

The saw palmetto plant or Serenoa repens (Bartr.) Small is a dwarf palm indigenous to the southeastern United States and West Indies.1 For many years saw palmetto extract has been used to treat symptomatic benign prostatic hyperplasia (BPH) throughout Europe.2, 3 However, available information on saw palmetto extract consists largely of in vitro experiments and European trials, of which many have limited value, as noted by Lowe et al.4, 5 Thus, we performed a randomized, placebo controlled trial to evaluate the tissue and clinical effects (primary and secondary outcome measures, respectively) of a saw palmetto herbal blend in men with symptomatic BPH.

MATERIALS AND METHODS

Included in our study were 44 ambulatory men 45 to 80 years old with symptomatic BPH but in good general health who were recruited from a pool of 160 eligible candidates treated at a general urology practice in metropolitan Los Angeles, California, between September 1997 and January 1998 (table 1). Appendix 1 shows study inclusion and exclusion criteria. All subjects signed a consent form approved by the Western Institutional Review Board, Seattle, Washington. Of the 44 randomized men 43 completed the study,

including 41 for whom complete data are available, while 1 died of myocardial infarction 4 months after randomization. We performed a 6-month randomized, double-blind placebo controlled trial (fig. 1 and table 2). After the 6-month visit all patients were invited to enter a long-term open label trial extension. We used a similar study design 2 years earlier for evaluation of finasteride.6, 7 A table of random numbers was used to assign subjects at a ratio of 1:1 to receive a sealed opaque bottle of soft gelatin capsules of a saw palmetto herbal blend or sealed opaque bottle of inert placebo capsules of glycerin, yellow wax, lecithin and a small quantity of soybean oil. To our subjects and us placebo was indistinguishable from the saw palmetto herbal blend capsule. Each preparation was ingested 3 times daily with meals. The herbal blend contained a standard preparation of 106 mg. of saw palmetto berry lipoidal extract and other herbs (table 3). We used a single product lot throughout the study. Average compliance, assessed by counting the remaining capsules at each visit, was 94.2% for all men during the trial. No subject had a compliance of less than 75% at any time.

Clinical outcomes included the International Prostate Symptom Score (I-PSS) and sexual function questionnaires, quality of life index, uroflowmetry, ultrasound of post-void residual urine volume, Tandem-R† prostate specific antigen (PSA) blood test, and testosterone, dihydrotestosterone and estradiol measurements. A panel of 26 conventional blood tests was also performed to obtain evidence of adverse effects (Appendix 2). Prostate volume was determined by magnetic resonance imaging (MRI), which permitted the measurement of whole and inner prostate volume separately using the formula for an ellipsoid. Sextant biopsy of the prostate was performed according to the method of Hodge et al.8 Each biopsy core was stained with india ink at the deep end to identify the transition zone. Detailed methodology of these outcome measures has been reported previously.6 Prostate biopsy cores were fixed in formalin and embedded in paraffin. Sections (5 m) were cut until tissue was exhausted, usually providing 15 to 20 slices. After routine histological study ruled out malignancy the remaining sections were preserved unstained for subsequent tissue analysis. After study completion all tissues were batch analyzed for evidence of atrophy, inflammation and neoplasia by hematoxylin and eosin staining by a single referee pathologist (J. I. E.), who remained blinded to clinical information. Atrophy and inflammation were estimated by 43 magnification assessment of the percent of atropic glands, and by scoring the overall core for the amount of acute and chronic inflammatory tissue. Estimates from all cores were then averaged. Prostate cores were also subjected to immunohistochemical staining and image analysis for quantitating tissue components (stroma, epithelium and lumen),6 apoptosis,9 cellular proliferation (Ki-67),9 angiogenesis (factor VIII),10 transforming growth factor-b11 and androgen receptor expression. 12 All data were entered into a personal computer via spreadsheet software and analyzed using a commercially available statistical package. As in a former study,6 statistical methods included the paired and 2-sample Student t tests, Kendall’s tau and linear regression with statistical significance considered at p ,0.05. Primary end points were the tissue changes that our study was designed and powered to detect. Clinical measures were secondary end points since previous analogous trials of finasteride showed that the study was not adequately powered to detect slight clinical

Clinical outcomes included the International Prostate Symptom Score (I-PSS) and sexual function questionnaires, quality of life index, uroflowmetry, ultrasound of post-void residual urine volume, Tandem-R† prostate specific antigen (PSA) blood test, and testosterone, dihydrotestosterone and estradiol measurements. A panel of 26 conventional blood tests was also performed to obtain evidence of adverse effects (Appendix 2). Prostate volume was determined by magnetic resonance imaging (MRI), which permitted the measurement of whole and inner prostate volume separately using the formula for an ellipsoid. Sextant biopsy of the prostate was performed according to the method of Hodge et al.8 Each biopsy core was stained with india ink at the deep end to identify the transition zone. Detailed methodology of these outcome measures has been reported previously.6 Prostate biopsy cores were fixed in formalin and embedded in paraffin. Sections (5 m) were cut until tissue was exhausted, usually providing 15 to 20 slices. After routine histological study ruled out malignancy the remaining sections were preserved unstained for subsequent tissue analysis. After study completion all tissues were batch analyzed for evidence of atrophy, inflammation and neoplasia by hematoxylin and eosin staining by a single referee pathologist (J. I. E.), who remained blinded to clinical information. Atrophy and inflammation were estimated by 43 magnification assessment of the percent of atropic glands, and by scoring the overall core for the amount of acute and chronic inflammatory tissue. Estimates from all cores were then averaged. Prostate cores were also subjected to immunohistochemical staining and image analysis for quantitating tissue components (stroma, epithelium and lumen),6 apoptosis,9 cellular proliferation (Ki-67),9 angiogenesis (factor VIII),10 transforming growth factor-b11 and androgen receptor expression. 12 All data were entered into a personal computer via spreadsheet software and analyzed using a commercially available statistical package. As in a former study,6 statistical methods included the paired and 2-sample Student t tests, Kendall’s tau and linear regression with statistical significance considered at p ,0.05. Primary end points were the tissue changes that our study was designed and powered to detect. Clinical measures were secondary end points since previous analogous trials of finasteride showed that the study was not adequately powered to detect slight clinical

RESULTS

At baseline the study and placebo groups were similar in demographic and clinical characteristics (table 1). Of the 44 men 41 finished the study with complete data available, 1 in the saw palmetto group died suddenly of unrelated cardiac causes halfway through the trial and 2 in the placebo group refused followup biopsy but otherwise finished the study with all other data complete. When the study was no longer blinded at the end of 6 months, all 23 placebo treated men entered the open label saw palmetto herbal blend extension. Of the 21 men receiving saw palmetto herbal blend 18 also elected to enter the extension. Table 4 shows the effects of intervention on symptom score, flow rate and post-void residual urine in the 2 groups. During the initial placebo lead-in month we noted a slight decrease in symptom score and increase in urinary flow in each group (not statistically significant). After randomization a slight progression of these trends continued. The ultimate change from baseline was slightly greater in the saw palmetto than in the placebo group (not statistically significant). No correlation was present between any of the clinical variables at baseline (table 1) and subsequent response to saw palmetto herbal blend.

Table 4 lists data on serum free-to-total PSA, testosterone, dihydrotestosterone and estradiol. Baseline serum levels of PSA were slightly higher in the placebo than in the saw palmetto herbal blend group (not significant). No statistically significant changes were noted in any of these measurements after saw palmetto herbal blend treatment. Table 4 also shows prostate volume at baseline and again after 6 months of randomization. No statistically significant volume change was apparent in the whole or inner prostate when measured separately.

In the saw palmetto group the atrophy index, defined as the percent of the prostate that was atrophic on 34 magnification inspection of the whole core, increased significantly after treatment from an average of 25.2% at baseline to 40.9% 6 months later (p ,0.01, fig. 2). In placebo treated patients the atrophy index was 43.1% at baseline, increasing only slightly to 48.4% after 6 months (not statistically significant). Figure 3 represents a sample of epithelial involution after saw palmetto herbal blend. No change in inflammatory features was observed in either group after treatment (data not shown). In the transition zone the percent epithelium decreased from a mean of 17.8% at baseline to 10.7% after 6 months of saw palmetto herbal blend therapy (p ,0.01, fig. 4). No significant change was noted in the stromal or luminal percent, or in any tissue component of the men treated with placebo. We detected a qualitatively similar change in peripheral zone epithelium before and after treatment but itwas not statistically significant (21.8% versus 17.7%). None of the other tissue studies, including those of apoptosis, angiogenesis, cellular proliferation, transforming growth factor-b or androgen receptor expression, had any statistically significant change after saw palmetto herbal blend. No serious adverse event was associated with saw palmettoherbal blend. A single episode of mild gastrointestinal upset involving cramps and diarrhea that may have been associated with treatment resolved spontaneously with continued use of the product. In each group 18 men were sexually active at baseline and only 1 had a decrease in ejaculatory volume that may have been due to therapy. No changes in libido or potency were reported. A panel of 26 standard blood tests revealed no changes from normal baseline values in either group (Appendix 2).

DISCUSSION

Plant based therapy is widely given in men with symptomatic BPH in western Europe, where physicians prescribe herbal products in the same manner as they prescribe drugs. For example, in Italy plant based products are prescribed 5 times more often than a-blocking agents or finasteride and in Germany more than 90% of all medications prescribed for symptomatic BPH involve phytotherapy.3 In a large multicenter trial Carraro et al observed that a French brand of saw palmetto extract was as effective as finasteride for decreasing symptoms and increasing urinary flow.13 In a recent meta-analysis of controlled European trials of saw palmetto extract Wilt et al noted that the saw palmetto effect significantly exceeded that of placebo in regard to overall urinary symptoms, nocturia and the maximal urinary flow rate.2 Thus, the interest in saw palmetto extract is great and its use in the United States is increasing rapidly,14 although to our knowledge controlled studies have not been performed in this country.

Lowe and Ku evaluated the European trials of saw palmetto extract and identified a number of shortcomings in many, including too few subjects, a lack of rigid inclusion and/or exclusion criteria, short duration, absence of product standardization and failure to administer a validated symptom questionnaire.4 Furthermore, in many studies no placebo effect was noted, raising additional concern since placebo mediated improvement was noted in virtually all clinical trials of BPH pharmaceuticals. Thus, Lowe and Ku stated that the available data may suggest but do not clearly establish the efficacy of saw palmetto extract for symptomatic BPH.4 A similar conclusion was drawn by Wilt et al, whose meta-analysis included more than 2,900 patients in 18 European trials.2 The lack of a statistically significant clinical benefit of saw palmetto herbal blend over placebo in our relatively small study groups is not surprising. In the largest saw palmetto extract data set of the meta-analysis of Wilt et al mean symptom score improvement over placebo was 1.4 points on a 19-point scale and uroflowmetry improved 1.9 cc per second.2 It is possible that in a larger series or one with longer followup the clinical differences in our 2 groups would become more pronounced.

The most important finding of our study is the morphometric observation that saw palmetto herbal blend causes involution of the prostatic epithelium in men with symptomatic BPH. Modest support for our morphometric finding was provided by histological studies, which revealed a treatment associated increase in the rate of atrophy (fig. 2). The percent of atrophy was different at baseline in the 2 groups. This finding was probably a chance result of our small sample size but it limits the strength of this supporting evidence. Histological data suggest but do not conclusively confirm our morphometric finding of epithelial involution. Morphometrically, involution was especially pronounced in the transition zone, where epithelial composition decreased from 17.8% at baseline to 10.7% after 6 months of saw palmetto herbal blend (p ,0.01). It is unclear how epithelial involution translates to clinical efficacy in the absence of prostate volume contraction. The epithelial contraction associated with saw palmetto herbal blend appears similar to that after finasteride treatment. 6 However, since saw palmetto treatment caused no decrease in serum PSA or prostate volume, the mechanism of action must be different. In fact, serum dihydrotestosterone did not decrease, indicating that the inhibition of 5-a reductase is not an important part of the mechanism (table 4).15–17 Since we noted no effect of saw palmetto herbal blend on prostate inflammatory tissues, we did not confirm the antiinflammatory hypothesis of Helpap et al.18 An unexplained finding in our series is the lack of a change in serum PSA despite marked epithelial contraction. This apparent paradox may involve some novel mechanism of action that was not detected in our studies of apoptosis, cell proliferation, growth factors, angiogenesis and androgen receptor expression. Our relatively small sample size may have caused skewing of the data in a way that obscured the true mechanisms. In this regard baseline atrophy data were different in the 2 groups, perhaps resulting from some degree of sampling error (fig. 4). Nevertheless, the involution of epithelial tissues confirmed by 2 independent and complementary

tests indicates that saw palmetto extract results in at least 1 measurable if currently unexplained effect on the human prostate. No change was noted in serum testosterone, dihydrotestosterone or estradiol, or in the prostate tissue expression of androgen receptors. In addition, sexual side effects were absent. Thus, the mechanism of action does not appear to be hormonal. In our study saw palmetto intervention involved an herbal blend and not pure saw palmetto extract (table 3). Thus, these data may or may not apply to pure saw palmetto

support the efficacy of any other single ingredient used alone for treating symptomatic BPH. However, nettle root extract and pumpkin seed oil have a long history of use in European preparations, often in combination with saw palmetto extract, for the treatment of BPH. We noted no adverse effects of saw palmetto herbal blend therapy. Patient desire for saw palmetto herbal blend treatment was great, as evidenced by the nearly uniform enrollment of subjects in each group into the open label extension. The interest of men in herbal remedies with a prostate condition was reported by Gerber et al in a study of BPH19 and by Nam et al in a study of prostate cancer.20 Due to the pathophysiological effect in our series and clinical efficacy in the large study of Wilt et al2 the saw palmetto option deserves consideration as first line intervention in men with symptomatic and uncomplicated BPH. However, current laws governing the manufacture and sale of herbal products in the United States do not protect consumers to the same extent as those governing pharmaceutical products.21–23 Hundreds of branded saw palmetto extract products are available. Thus, specific saw palmetto extract products as well as other phytotherapeutic substances should be evaluated within the medical mainstream and then, as appropriate, offered in perspective to the many patients who desire “safe, natural remedies.”

CONCLUSIONS

Saw palmetto herbal blend for symptomatic BPH resulted in the contraction of prostatic epithelial tissues, apparently via a nonhormonal mechanism. Serum PSA was unchanged. The effect on symptom score and urinary flow was mild but statistically significant in large studies. Thus, saw palmetto herbal blend appears to be a reasonable alternative for men with early, uncomplicated prostatism. No major side effects were observed.