SEARCH

SEARCH BY CITATION

Keywords:

  • acute urinary retention;
  • benign prostate hyperplasia;
  • transurethral resection of prostate;
  • postoperative complication

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES
  9. Supporting Information

Study Type – Prognosis (cohort)

Level of Evidence 2a

What's known on the subject? and What does the study add?

Higher mortality and morbidity rates in men presenting with AUR have been reported in previous studies.

This study has comprehensive comparisons of post-TURP complications between patients with and without AUR. Furthermore, it shows that AUR is associated with increased risk of complications after TURP.

OBJECTIVE

  • • 
    To assess the association between a history of acute urinary retention (AUR) and complications after transurethral resection of prostate (TURP).

PATIENTS AND METHODS

  • • 
    We conducted a retrospective, national, population-based study using Taiwan's National Health Insurance Research Database.
  • • 
    We included men > 50 years old, diagnosed with benign prostatic hyperplasia (BPH) and divided these into two groups: an AUR+ group – those with AUR who underwent TURP between 2002 and 2004; and an AUR- group – those without AUR who underwent TURP between those dates.
  • • 
    Prostate cancer, Parkinsonism and multiple sclerosis were exclusion criteria.
  • • 
    Postoperative complications, e.g. re-catheterization, haematuria or urinary tract infection (UTI), were compared using crude odds ratios (ORs), 95% confidence intervals (CIs), and Student's t-test.
  • • 
    A chi-squared test was used for potential confounding factors: preoperative UTI and anticoagulant use.
  • • 
    Univariate and multivariate analysis on medical expenses were conducted.

RESULTS

  • • 
    The AUR+ group contained 3305 men; the AUR- group contained 1062.
  • • 
    Re-catheterization (13.8%), septicaemia (1.1%) and shock (0.3%) were found only in the AUR+ group.
  • • 
    The AUR+ group had more UTIs (18.9% vs. 15.6%, OR: 1.26, 95% CI: 1.05–1.52), more lower urinary tract symptoms (22.8% vs. 16.9%, OR: 1.45, 95% CI: 1.21–1.73), fewer blood transfusions (3.2% vs. 1.5%, OR: 2.19, 95% CI: 1.29–3.72) and higher medical expenses.
  • • 
    There were no significant differences in haematuria, lower urinary tract stricture, or re-surgical intervention of the prostate and second-line antibiotic use.

CONCLUSION

  • • 
    Patients in Taiwan with BPH with AUR who were treated by TURP were associated with a higher risk of complications, longer hospital stay and more comorbidities than those without AUR and a preoperative warning is warranted for these patients.

Abbreviations
NHIRD

National Health Insurance Research Database

ICD-9-CM

International Classification of Diseases, 9th revision, Clinical Modification

OR

odds ratio

NT$

New Taiwan dollars.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES
  9. Supporting Information

Acute urinary retention is a severe symptom of men with BPH. It is defined as a sudden and painful inability to void voluntarily [1,2]. Although the causes of AUR are variable, BPH accounts for most episodes. The prevalence rate of AUR in men with BPH is estimated to be as high as 53% [3]. AUR is not only painful, but also the harbinger of severe systemic diseases. Higher mortality and morbidity rates in men presenting with AUR have been reported in previous studies [4,5].

In Western countries, AUR was the chief compliant in 20–42% of men who underwent TURP [6]. Escalating postoperative complications and longer hospital stays in men with BPH who develop AUR have been reported [5–7]. Patients who presented with AUR had a high mortality rate in the first 3 years after prostatectomy [8].

Most of the AUR-related complications reported in previous studies, however, were not primary endpoints in their study settings. Comprehensive comparisons of post-TURP complications between patients with and without AUR are lacking. Furthermore, these studies did not prove that patients given elective TURP did not have a history of AUR. The present population-based study provides a general view on the differences of post-TURP complications between patients with AUR (AUR+) and those without (AUR-) in Taiwan.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES
  9. Supporting Information

STUDY POPULATION AND DATA COLLECTION

This retrospective, population-based study was designed to investigate complications in patients with AUR who underwent TURP in Taiwan between 1 January 2002 and 31 December 2004. The data were extracted from the non-sampled National Health Insurance Research Database (NHIRD), which were obtained from the National Health Research Institutes of Taiwan. The NHIRD was implemented in March 1995 and covered almost 97% of the population in Taiwan at the end of 2002 [9]. In the NHIRD, we were able to retrieve records of clinic visits, hospital admissions, prescriptions, and medical procedures received, as well as demographic information (e.g. age, sex, residence), while personal identification of each patient was encrypted and anonymized. Also, each record contained diagnoses and procedures coded according to the International Classification of Diseases, 9th revision, Clinical Modification ([ICD-9-CM] Appendix).

DEFINITIONS

To select the primary AUR cases, we allowed a 6-month period without any history of Foley catheterization before the first record of AUR appeared in the study period. An earlier study [10] has indicated that 80% of AUR cases tended to recur within 6 months. AUR was considered established if it was the primary diagnosis, or if BPH appeared with a Foley catheterization procedure. LUTS included urinary incontinence, dysuria, urinary frequency and others (e.g. splitting of urinary stream, slowing of urinary stream). These symptoms were all identifiable in the ICD-9-CM. Lower urinary tract strictures included meatal, urethral, and bladder neck strictures.

STUDY DESIGN AND INCLUSION CRITERIA

Men who were >50 years old, diagnosed with BPH, and given TURP were all included. The patients were divided into two groups based on whether or not they had AUR. All the patients in the AUR+ group conformed to the event sequence: diagnosed with BPH, compliant with the criteria of AUR, and then received TURP. Patients were excluded if they had prostatic cancer, Parkinsonism, or multiple sclerosis before TURP. We compared post-TURP complications between the two groups.

Preoperative information, including age, urinary retention episodes, Charlson comorbidity index (CCI) score, UTIs, LUTS, lower urinary tract strictures, and anticoagulant (aspirin, warfarin and clopidogrel) use, was analysed. During admission for TURP, we focused on septicaemia, shock, blood transfusion, second-line antibiotic use, length of hospital stay, and medical expenses. In Taiwan, first- or second-generation cephalosporins are the most commonly prescribed post-TURP antibiotics. When third- or fourth-generation cephalosporins are used, postoperative infection is usually serious. At 2 weeks after surgery, Foley catheterization, haematuria and UTI events were followed-up. LUTS, lower urinary tract strictures, and prostate reoperation were traced until the end of the study. Patients with each of the above diagnoses were included once the diagnoses were established. Patients with lower urinary tract strictures were included only when they had been given surgical interventions for strictures. That usually meant strictures that were more severe.

UNDERLYING COMORBIDITY

Because no illness severity index is currently available in Taiwan, we used the CCI score to adjust for patients' comorbidities. The CCI score, developed in 1987 [11], assesses the risk of death from comorbidities. It has been adapted according to ICD-9-CM codes and is widely used in administrative databases [12]. All the comorbidities included in the CCI were searched before or during admission for TURP. Each patient was evaluated for at least 6 months.

STATISTICAL ANALYSIS

For the comparisons of complications in different groups, the crude odds ratio (OR) and the 95% CI were estimated. We used Student's t-test to assess each complication. All statistical tests were two-sided, and a P value < 0.05 was considered to indicate statistical significance. For the associations between outcomes and potential confounding factors, a chi-squared test, and univariate and multivariate logistic regression analysis were used. Univariate and multivariate linear regression analyses of medical expenses were carried out. sas/stat 9.2 and pasw statistics (formerly spss) 18 were used for all statistical analyses.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES
  9. Supporting Information

From the NHIRD, we identified 331 188 men who had a BPH diagnosis between 1 January 2002 and 31 December 2004. Approximately 3.3% of them (n= 11 036) underwent TURP. Patients with prostate cancer (n= 402), Parkinsonism (n= 87) and multiple sclerosis (n= 1) were excluded. In addition, 6179 men were excluded because they did not fit the criteria of the definition of AUR or meet the event sequence. Finally, 3305 patients (mean [sd] age 72.3 [8.0] years) were included in the AUR+ group, and 1062 patients (mean [sd] age: 70.4 [7.8] years) without any AUR episodes in the AUR- group (Table 1). In the AUR+ group, 2373 patients (71.8%) underwent TURP < 1 month after an AUR episode, and 2618 patients (79.2%) < 3 months after an AUR episode.

Table 1. Demographic data of the AUR+ and AUR- groups
VariableAUR+ group, N= 3305AUR- group, N= 1062 P
Mean (sd; range), years72.3 (8.0; 50–94)70.4 (7.8; 50–97)0.648
Mean no. of AUR episodes2.60NA
Mean (sd) CCI score2.9 (2.5)2.5 (2.1)<0.001
Age distribution, years, n (%)  <0.001
50–59219 (6.6)85 (8.0)
60–69935 (28.3)387 (36.4)
70–791539 (46.6)471 (44.4)
80–89574 (17.4)115 (10.8)
≥9038 (1.1)4 (0.4)

There were no post-TURP Foley catheterizations in the AUR- group (Table 2). In the AUR+ group, the mean episodes of preoperative AUR were 2.6 (Table 1). A total of 13.8% of patients in the AUR+ group needed to be re-catheterized after surgery (Table 2). In this subgroup of patients, the mean episodes of preoperative AUR were 4.0, higher than that of the AUR+ group as a whole.

Table 2. Comparison of complications between the AUR+ and AUR−− groups
VariablesAUR+ group, N= 3305AUR- group N= 1062OR P
  • *

    Splitting of urinary stream, slowing of urinary stream.

Foley, n (%)457 (13.8)0NA<0.001
Haematuria269 (8.1)79 (7.4)1.10 (0.85–1.43)0.46
UTI626 (18.9)166 (15.6)1.26 (1.05–1.52)0.01
Lower urinary tract stricture87 (2.6)34 (3.2)0.82 (0.55–1.22)0.32
Meatal2 (2.3)0
Urethral29 (33.3)16 (47.1)
Bladder neck56 (64.4)18 (52.9)
Re-surgery of prostate33 (1.0)14 (1.3)0.76 (0.40–1.42)0.38
LUTS754 (22.8)180 (16.9)1.45 (1.21–1.73)<0.001
Incontinence244 (7.482 (7.7)
Frequency300 (9.1)68 (6.4)
Dysuria201 (6.1)28 (2.6)
Other*9 (0.3)2 (0.2)
Septicaemia25 (1.1)0NA<0.001
Shock10 (0.3)0NA<0.001
Blood transfusion107 (3.2)16 (1.5)2.19 (1.29–3.72)0.004
Antibiotics29 (0.9)7 (0.7)1.33 (0.58–3.05)0.49
Mean (sd) length of hospital stay, days6.4 (7.2)4.6 (2.0) <0.001
Mean (sd) medical expense44 677.5 (32 690.3)38 070.0 (8567.1) <0.001

The differences in haematuria < 2 weeks after TURP were not significant. Preoperative anticoagulant use in the AUR+ and AUR- groups was examined. The correlation in each group between anticoagulant use and postoperative haematuria was not significantly different between the groups (Table 3). The rate of postoperative UTI was higher in the AUR+ group (18.9% vs. 15.6%, P= 0.01; OR: 1.26, 95% CI 1.05–1.52). No patients in the AUR- group had a preoperative UTI; however, 805 patients (24.4%) in the AUR+ group had preoperative UTIs, and 161 patients had postoperative UTIs. Of 2500 patients without preoperative UTIs, 469 (18.8%) developed postoperative UTIs. The null hypothesis could not be excluded (P= 0.44). This result suggested a greater number of ascending UTIs in men with Foley catheterization.

Table 3. Anticoagulant use among patients with haematuria and patients given blood transfusions
GroupAnticoagulant use (+), nAnticoagulant use (−), n P
AUR+
Haematuria (+)432280.27
Haematuria (−)4092625 
Blood transfusion (+)20870.14
Blood transfusion (−)4322766 
AUR-
Haematuria (+)12670.22
Haematuria (−)105878 
Blood transfusion (+)4150.16
Blood transfusion (−)113930 

The rates of both lower urinary tract stricture (2.6% vs. 3.2%, P= 0.32) and prostate re-operation (1.0% vs. 1.3%, P= 0.38) were slightly lower in the AUR+ group than in the AUR- group. Postoperative LUTS were more prevalent in the AUR+ group than in the AUR- group (22.8% vs. 16.9%, OR = 1.45, 95% CI: 1.21–01.73). In both groups, incontinence and frequency accounted for >70% of LUTS. In the AUR+ group, frequency was the most common lower urinary tract symptom (39.8%), while in the AUR- group, incontinence was (45.6%).

Septicaemia (1.1%) and shock (0.3%) during hospitalization occurred only in the AUR+ group. The blood transfusion rate was twice as high in the AUR+ group as in the AUR- group (3.2% vs. 1.5%, OR = 2.19, 95% CI: 1.29–3.72). Preoperative anticoagulant use had no significant effect on blood transfusion (Table 3), nor was there a significant difference in the use of second-line antibiotics. The mean (sd) length of hospital stay was 6.4 (7.2) days in the AUR+ group and 4.6 (2.0) days in the AUR- group (P < 0.001), and medical expenses were New Taiwan dollars (NT$) 44 677.5 (32 690.3) in the AUR+ group and NT$38 070.0 (8567.1) in the AUR- group (P < 0.001), which suggested more in-hospital complications in the AUR+ group.

Univariate analysis showed that age, hospitalization stays (regardless of whether or not the patient had AUR, and CCI score were all significantly correlated with medical expenses (Table 4). The mean medical expense was NT$6607.5 higher in the AUR+ group than in the AUR- group. In multivariate analysis, the length of hospital stay and the CCI score remained significant.

Table 4. Univariate and multivariate analysis of medical expenses
 UnivariateMultivariate
B estimate P B estimate P
Age342.0<0.00162.90.109
Length of hospital stay3215.2<0.0013182.7<0.001
CCI score1711.8<0.001324.80.012
Group (AUR+ or AUR-)6607.5<0.001487.70.502

Both the univariate and multivariate analysis (Table 5) showed that the group factor was associated with re-catheterization and postoperative UTI. For the lower urinary tract stricture and second-line antibiotics, the group factor was more persuasive than CCI score factor in multivariate analysis. Septicaemia, shock, blood transfusion, length of hospital stay and LUTS were all correlated with group factors and CCI score, but in group factors especially, the ORs were much higher.

Table 5. Univariate and multivariate analysis of complications
VariablesGroup (AUR+ or AUR-)
UnivariateMultivariate
OR (95%CI) P OR (95%CI) P
  1. Age was corrected in multivariate analysis.

FoleyNA<0.001NANA
Haematuria1.10 (0.85–1.43)0.461.17 (0.90–1.52)0.24
UTI1.26 (1.05–1.52)0.011.63 (1.35–1.98)<0.001
Lower urinary tract stricture0.82 (0.55–1.22)0.320.54 (0.30–0.99)0.05
Re-surgery of prostate0.76 (0.40–1.42)0.380.71 (0.37–1.33)0.29
LUTS1.45 (1.21–1.73)<0.0011.40 (1.15–1.69)0.001
SepticaemiaNA<0.001NANA
ShockNA<0.001NANA
Blood transfusion2.19 (1.29–3.72)0.0041.92 (1.13–3.27)0.02
Antibiotics1.33 (0.58–3.05)0.498.14 (1.1–60.0)0.04
Hospitalization <0.001 <0.001
Variables CCI score
Univariate Multivariate
OR (95%CI) P OR (95%CI) P
Foley1.03 (0.99–1.08)0.101.02 (0.98–1.06)0.44
Haematuria0.98 (0.94–1.03)0.400.99 (0.94–1.03)0.57
UTI0.98 (0.95–1.01)0.250.99 (0.95–1.02)0.41
Lower urinary tract stricture1.07 (1.00–1.15)0.050.19 (0.97–1.20)0.19
Re-surgery of prostate1.12 (1.01–1.23)0.031.12 (1.01–1.24)0.03
LUTS1.04 (1.01–1.07)0.021.04 (1.01–1.07)0.02
Septicaemia1.32 (1.19–1.47)<0.0011.33 (1.19–1.48)<0.001
Shock1.27 (1.06–1.51)0.011.23 (1.02–1.47)0.03
Blood transfusion1.11 (1.04–1.18)0.0021.08 (1.00–1.15)0.03
Antibiotics1.16 (1.03–1.3)0.021.12 (1.00–1.27)0.06
Length of hospital stay <0.001 <0.001

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES
  9. Supporting Information

To our knowledge, this is the first large-scale population-based cohort study in Asia to compare AUR complications after TURP. To avoid possible confounding factors, we excluded patients diagnosed with prostate cancer, Parkinsonism, and multiple sclerosis. Only men meeting the AUR criteria were included. CCI score, preoperative anticoagulant use, and UTIs were investigated and analysed using a chi-squared test. We found that re-catheterization, in-hospital septicaemia and shock occurred only in the AUR+ group. Patients in the AUR+ group developed more postoperative UTIs and LUTS, needed more blood transfusions, had longer hospital stays, and had higher medical expenses than did patients in the AUR- group. We thus inferred that patients with BPH with AUR have a higher risk of complications after TURP.

In the present study, 32.6% of the patients had AUR before TURP. According to earlier studies, AUR accounted for 27%, 23%, and 43% of all patients with BPH undergoing prostatectomy in North America, the UK and Singapore, respectively [5,6,13]. In the Singapore report, the trend declined decennially from 79 to 43%. The decreasing rate may be attributed to the advancement of medical therapy or early surgical interventions.

Interestingly, we also found that no one in the AUR- group had been re-catheterized after surgery. This finding suggested that the detrusor contractile function of men with AUR might have been injured after urinary retention [14]. In earlier studies, the reported failure to void rate after TURP ranged from 3.0 to 12% [5,15–20]. The present findings suggested a higher rate. A possible explanation is that most previous studies did not separate patients into AUR+ and AUR- groups. The present study showed there was a tendency for re-catheterization after TURP in patients in the AUR+ group. These patients should be informed about the possibility of re-catheterization after TURP.

The postoperative haematuria rate was slightly higher in the AUR+ group, but not significantly so. Elevated postoperative UTIs and re-catheterization rate in the AUR+ group may delay wound healing and make the tissue more fragile. The haemostasis procedure may also affect the result. Preoperative anticoagulant use was not associated with haematuria. We inferred that most surgeons would suggest that patients stop anticoagulant use 7 days before TURP to avoid surgical bleeding. In addition, because our data came from the NHIRD, we were unable to examine the patients' drug compliance.

The postoperative UTI rate in the AUR+ group was 18.9%, higher than in the AUR- group. In Pickard et al. [6], the total urinary infection rate (including suspected and proved UTI, and epididymo-orchitis) was 18.6%. Nonetheless, the overall post-TURP UTI rate reported in the literature was <5%, except for 25% in Doll et al. [15]. Both Pickard et al. [6] and the present study found more prevalent infection events in the AUR+ group. The reason for this discrepancy may be different study definitions of UTI. The NHIRD was established as an administrative tool; detailed clinical laboratory data, such as urine analyses and cultures, were not available for the present study. In addition, a higher re-catheterization rate in the AUR+ group may have led to a greater prevalence of ascending UTIs.

In the literature [21], the urethral stricture rate ranges from 2.2 to 9.8% and the bladder neck stricture rate from 0.3 to 9.2%. In the present study, the lower urinary tract stricture rates in the two groups were ∼3%. Theoretically, more Foley catheterizations and the greater number of UTIs in the AUR+ group would lead to a higher urinary tract stricture rate, but there were no differences in lower urinary tract stricture rates between the two groups. Surgical technique, choice of instrument size, and sufficient or insufficient lubricant may all account for the background biases [21]. TURP was a definitive treatment for BPH. The retreatment rate of TURP has been reported to be 3–14.5% after 5 years [21]. In the present study, the mean time before re-surgery of the prostate was 62 days in the AUR+ group and 81 days in the AUR- group (data not shown). Reasons for an early return to the operation room included a blood clot tamponade, uncontrolled bleeding, or an insufficient resection. There was no significant difference in the re-surgery rate between the two groups, but the AUR+ group returned to the operation room earlier than the AUR- group.

In the present study, there were a greater number of LUTS complaints in the AUR+ group. These complaints were subjective. In addition to vesical aetiology, the greater number of postoperative re-catheterizations and UTI episodes in the AUR+ group may explain the discrepancy between the two groups. Incontinence rates were reported as 0.3% in Haupt et al. [17], 1.0% in Kuntz et al. [16], and 38% in Doll et al. [15]. The different definition of LUTS used in the present study may account for the discrepancy between incontinence rates and the aforementioned studies. The present study included not only incontinence, but also frequency, dysuria, and others (e.g. splitting of urinary stream and slowing of urinary stream). Although the IPSS has provided a reliable tool for assessing LUTS, the information needed was not available in the NHIRD. Although LUTS were common during the early postoperative period, one study [22] followed-up the IPSS 1 year after TURP and discovered that the IPSS gradually decreased (improved) from 25 to 4 points. Obviously, the AUR had some impact on urinary bladder function or perception. Although it was not a lethal condition, it contributed to a deterioration in quality of life.

In the present study, there were more in-hospital complications in the AUR+ group, including septicaemia, shock, second-line antibiotic use and blood transfusions. The first two events did not even occur in the AUR- group, and the incidence rates were not higher than in previous reports. The blood transfusion rate in the AUR+ group was twice that in the AUR- group. The larger prostate in AUR+ men may explain that difference. Prostatic infarction may also have been associated with it. Prostatic infarction has long been suspected to be one of the aetiologies of AUR [23,24]. Post-infarction fragile vessel integrity may lead to much more postoperative bleeding. Further histopathological study may help to clarify whether there is an association. Finally, the longer hospital days and higher medical expenses both suggested a more complicated course of hospitalization for the AUR+ group. The greater comorbidity in AUR+ group patients was associated with higher medical expenses both in univariate and in multivariate analysis. Thus, we suggest that BPH patients with AUR should be treated as a special group. A preoperative warning and multidisciplinary medical care are warranted for these patients.

The present study has some limitations. First, the NHIRD lacked some clinical information, such as prostate size and IPSS, which is a universal methodological limitation when using administrative databases; however, BPH is a clinical diagnosis and the symptoms may not always be proportional to prostate size. For most patients, the obstructive symptoms were resolved after TURP. Most patients complained of irritating symptoms, such as incontinence and urinary frequency and these patients would have higher IPSS scores. Second, the effect on AUR of medications such as anticholinergics was not excluded in our study as we did not exclude patients taking these medications. This did not affect our results significantly; the incidence rate of medication related to AUR was as low as 2% [3,25]. In addition, most of the patients susceptible to the medication effect already had clinical BPH symptoms. Third, second-line antibiotics are prescribed only in the hospital and under strict regulations by the infection control centre of each hospital in Taiwan. We believe that the record of antibiotic use is accurate.

In conclusion, the present study is a real-world study reflecting current practice in Taiwan. Our findings support the hypothesis that BPH patients with AUR who undergo TURP are associated with a higher risk of complications than those without. Re-catheterization, postoperative UTIs, LUTS, septicaemia, shock, blood transfusions, second-line antibiotic use, hospitalization days, and medical expenses were all higher in the AUR group. The underlying mechanism of AUR on the post-TURP complications is not clear. Further pathophysiological studies are necessary to elucidate this. Preventing AUR secondary to BPH implies delaying the progression in patients at risk and improving the prognosis when they are given TURP. Once AUR develops, a preoperative warning and cautious postoperative care are warranted for these patients.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES
  9. Supporting Information

APPENDIX: ICD-9-CM coding used in the present study

FilenameFormatSizeDescription
bju11471_sm_Appendix.doc34KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.