Long-term outcomes after surgical and nonsurgical management of chronic pelvic pain: one year after evaluation in a pelvic pain specialty clinic.
ABSTRACT The purpose of this study was to describe long-term outcomes for women with chronic pelvic pain (CPP) after evaluation in a CPP specialty clinic.
This was a prospective observational cohort study of women treated for CPP at the UNC Pelvic Pain clinic between 1993 and 2000. The primary outcome was improvement in pain and the main exposure was treatment group: primarily medical (pharmacotherapy, psychotherapy, physical therapy, or combinations of the 3) or surgical (hysterectomy, resection or ablative procedures, oophrectomy, diagnostic surgery, pain mapping, vulvar or vestibular repair). Univariate, bivariate, and multivariable analyses were performed to look for relationships between background characteristics, treatment group, and improvement in pain.
Of 370 participants; 189 had surgical treatment and 181 had medical treatment. One year after evaluation, 46% reported improvement in pain and 32% improvement in depression. Improvement in pain was similar in both treatment groups and odds of improvement were equal even after adjusting for background characteristics, psychosocial comorbidity, and previous treatments.
One year after evaluation in a CPP specialty clinic, women experienced modest improvements in pain and depression after recommended surgical or nonsurgical treatment.
- SourceAvailable from: Jørgen Nordling[Show abstract] [Hide abstract]
ABSTRACT: Background / Purpose: Our aim here was to examine the evidence for an effect of physiotherapy on pain, physical activity and quality of life in the treatment of female chronic pelvic pain (CPP). Main conclusion: Narrative synthesis of the results, based on effect estimates and clinically relevant pain improvement, disclosed some evidence to support an effect of multidisciplinary intervention and Mensendieck somatocognitive therapy on female CPP.European Journal of Pain Supplements 09/2011; 5(1):267-268.
- Revista colombiana de obstetricia y ginecología 12/2010; 61(4):329-334.
- [Show abstract] [Hide abstract]
ABSTRACT: Most chronic non-cancer pain (CNCP) conditions are more common in women and have been reported to worsen particularly during the peak reproductive years. This phenomenon suggests that ovarian hormones might play a role in modulating CNCP pain. To this end, we reviewed human literature aiming to assess the potential role of ovarian hormones in modulating the following CNCP conditions: musculoskeletal pain, migraine headache, temporal mandibular disorder, and pelvic pain. We found 50 relevant clinical studies, the majority of which demonstrated a correlation between hormone changes or treatments and pain intensity, threshold, or symptoms. Taken together, the findings suggest that changes in hormonal levels may well play a role in modulating the severity of CNCP conditions. However, the lack of consistency in study design, methodology, and interpretation of menstrual cycle phases impedes comparison between the studies. Thus, while the literature is highly suggestive of the role of ovarian hormones in modulating CNCP conditions, serious confounds impede a definitive understanding for most conditions except Menstrual Migraine and endometriosis. It may be that these inconsistencies and the resulting lack of clarity have contributed to the failure of hormonal effects being translated into medical practice for treatment of CNCP conditions.Pain 08/2014; · 5.64 Impact Factor
Long-term outcomes after surgical and nonsurgical
management of chronic pelvic pain: One year after
evaluation in a pelvic pain specialty clinic
Georgine Lamvu, MD, MPH,a,* Rachel Williams, PhD,bDenniz Zolnoun, MD, MPH,a
Mary E. Wechter, MD,aAnne Shortliffe, RN,aGrace Fulton, BS,aJohn F. Steege, MDa
Division of Advanced Laparoscopy and Pelvic Pain, Department of Obstetrics and Gynecology,aUniversity of North
Carolina School of Medicine, Chapel Hill, NC; and Department of Epidemiology,bUniversity of North Carolina
School of Public Health, Chapel Hill, NC
Received for publication December 15, 2005; revised March 2, 2006; accepted March 19, 2006
Chronic pelvic pain
Objective: The purpose of this study was to describe long-term outcomes for women with chronic
pelvic pain (CPP) after evaluation in a CPP specialty clinic.
Study design: This was a prospective observational cohort study of women treated for CPP at the
UNC Pelvic Pain clinic between 1993 and 2000. The primary outcome was improvement in pain
and the main exposure was treatment group: primarily medical (pharmacotherapy, psychother-
apy, physical therapy, or combinations of the 3) or surgical (hysterectomy, resection or ablative
procedures, oophrectomy, diagnostic surgery, pain mapping, vulvar or vestibular repair). Univar-
iate, bivariate, and multivariable analyses were performed to look for relationships between back-
ground characteristics, treatment group, and improvement in pain.
Results: Of 370 participants; 189 had surgical treatment and 181 had medical treatment. One
year after evaluation, 46% reported improvement in pain and 32% improvement in depression.
Improvement in pain was similar in both treatment groups and odds of improvement were equal
even after adjusting for background characteristics, psychosocial comorbidity, and previous
Conclusion: One year after evaluation in a CPP specialty clinic, women experienced modest
improvements in pain and depression after recommended surgical or nonsurgical treatment.
? 2006 Mosby, Inc. All rights reserved.
Rachel E. Williams presently works at Glaxo SmithKline (GSK) and owns stock in GSK, but her participation in this study occurred when she
was affiliated with UNC and before she began working for GSK.
Presented at the Sixty-Eighth Annual Meeting of the South Atlantic Association of Obstetricians and Gynecologists, Orlando, FL, January
29-February 1, 2006.
* Reprint requests: Dr. Georgine Lamvu, Department of Obstetrics and Gynecology, University of North Carolina at Chapel Hill, 101 Manning
Drive, CB 7570 MacNider Building, Chapel Hill, NC 27599-7508.
0002-9378/$ - see front matter ? 2006 Mosby, Inc. All rights reserved.
American Journal of Obstetrics and Gynecology (2006) 195, 591–600
Chronic pelvic pain (CPP) is a poorly defined disor-
der that affects approximately 15% of women.1It is the
primary indication for an estimated 40% of diagnostic
laparoscopies, 10% of hysterectomies, and it accounts
for more than 2 billion dollars in health care costs annu-
ally.1,2In spite of the high prevalence and health care
costs associated with this disorder, CPP remains poorly
understood and largely understudied.
Commonly, CPP is defined as noncyclic pain of at
least 6 months duration, localized to the pelvis, anterior
abdominal wall, at or below the umbilicus and lower
back and buttocks. Pain levels are quantified as severe
enough to cause disability and require medical care but
are often unresponsive to treatment.2-6CPP may arise
from 1 or more organ systems and is often associated
with psychologic disturbances (depression, anxiety), a
history of sexual and physical abuse, and a variety of
somatic complaints.7This complexity makes patients
with CPP difficult to evaluate, diagnose, and treat.
Treatment options for women with CPP vary from
medical to surgical interventions. Surgical treatments
range from diagnostic laparoscopy, resection of endo-
metriosis, and neuroablative procedures, to partial resec-
tion of pelvic organs or full hysterectomy.7,8Medical
treatments commonly include hormonal suppression
of the ovarian cycle, pain medications, physical therapy,
and cognitive-behavioral therapy or psychotherapy.
Multidisciplinary diagnostic and management strategies
long-term outcomes of treatment in women with CPP.
Our goals were to: (1) describe long-term outcomes,
undergoing multidisciplinary treatment in a CPP clinic;
(2) determine if a variety of psychosocial patient charac-
teristics are associated with outcomes; and (3) determine
whether there is an association between improvement in
pain levels and the type of primary treatment received.
Material and methods
Study setting and study population
Between June 1993 and December 2000, the UNC Pelvic
Pain Clinic evaluated 987 women referred predominately
from North Carolina. Patient evaluation and data col-
lection methods have been previously published.9,10The
UNC Institutional Review Board approved the use of
clinical data for research purposes.
written questionnaires including a general medical infor-
mation form, a history of physical abuse survey, the Beck
Depression Inventory, and the McGill Pain Question-
naire.11,12The 3 physicians involved collected information
on previous surgeries, current and past treatments,
and performed a complete examination. Their findings,
diagnostic impressions, and recommended treatments
(medications, surgeries, referrals), as well as subsequent
treatments, surgeries, and pathology diagnoses were all
recorded. Recommended treatments included surgical,
medical (pain management, physical therapy, or psycho-
therapy), or combination therapy. One year after the
initial evaluation, patients were again asked to complete
the McGill Pain Questionnaire and the Beck Depression
Inventory; 370 women completed these questionnaires.
Patient information was abstracted from standardized
clinical history forms, questionnaires, and patient med-
ical records and entered into a SAS database (SAS,
Data collection was limited to women with self-
reported pelvic pain lasting 6 months or longer, al-
though it was possible for a patient to be referred for
pelvic pain lasting less than 6 months.9,10The pain was
usually localized to 1 or more of the following anatomic
areas: abdomen (below the umbilicus), pelvic organs
(ovaries, uterus, urethra, bladder, pelvic floor muscles),
lower back, vulva, or vagina. Clinical diagnoses were
grouped as ‘‘yes’’ if the physician indicated that the diag-
nosis was definite or probable and ‘‘no’’ if the physician
indicated that the diagnosis was not probable, did not
apply to the patient, or was only ‘‘possible.’’ If neces-
sary, the diagnoses were confirmed using records from
subsequent visits. Clinical diagnoses included muscular
back pain, pelvic floor tension myalgia, endometriosis,
pyriformis syndrome, vaginismus, pelvic congestion syn-
drome, myofascial syndrome, adhesions, adenomyosis,
fibroids, urethral syndrome, pelvic floor relaxation, vul-
vodynia, and vulvar vestibulitis syndrome. Information
on previous diagnoses and surgeries was collected in a
similar manner. Thus,thetotalnumberofprevioussurger-
ies was generated based on medical record, patient report,
and confirmation that the surgery was performed for the
same pain that ended in referral to the UNC CPP clinic.
Pain location was described by patients and in-
cluded: left, right, or middle pelvic area; low back;
vaginal opening; deep vagina; high in the pelvis near the
navel or in the thighs. These locations were grouped to
create the variable ‘‘total number of pain sites’’ because
many women reported more than 1 location of pain.
Pain levels were measured using the McGill Pain
Questionnaire (MPQ), a validated written self-question-
naire for measuring sensory and affective components
History of abuse included adult or child sexual abuse,
rape at any age, adult physical abuse, or traumatic
discipline during childhood. Depression was measured
with the Beck Depression Inventory (BDI), which clas-
sifies patients as not depressed, minimally depressed,
moderately depressed, and severely depressed.11
592Lamvu et al
Additional background characteristics collected for
each patient included age, self-described race, income,
educational level, and marital status.
Medical treatments included analgesics (opioid and
hypnotic/anticonvulsant, estrogens, progestins, combi-
nation estrogens and progestins, injectable progestins,
gonadotropin releasing hormone agonists, nonsteroidal
anti-inflammatory drugs, trigger point injections, phys-
ical therapy, and psychotherapy in various combina-
tions. Management of pain medications and trigger
point injections was performed by our clinic physicians,
while physical therapy and psychologically based ther-
apies were conducted by certified therapists within our
referral network. Surgical treatments ranged from min-
imally diagnostic laparoscopy to lysis of adhesions,
excision or ablation of endometriosis, unilateral or
bilateral oophrectomy, ovarian cystectomy, pain map-
ping, uterine suspension, uterosacral ablation, and hys-
terectomy. Laparotomy, vaginal, and vulvar surgery
was also included in the surgical treatment group.
For the purposes of this study, patients were categorized
into 2 exposure groups: medical or surgical treatment.
Medical treatment patients may have had previous
medical or surgical therapy before referral, but did not
receive surgical therapy while at UNC. The surgical
group included all women who had surgery at UNC,
regardless of previous treatment. Patients who received
both medical and surgical therapy at UNC were con-
sidered part of the surgical study group.
The main outcome of our analysis was improvement
in pain 1 year after initial evaluation. Pain levels were
grouped into the MPQ total score categories: no pain
or minimal pain (MPQ score 0-15); mild pain (MPQ
score 16-25); moderate pain (MPQ score 26-35); and
severe pain (MPQ score R36). BDI depression scores at
baseline and at 1 year were grouped as follows: no
depression (0-9); mild or minimal (10-18); moderate
(19-29); and severe depression (BDI score R30).
level from baseline to follow-up was defined as: (1) wors-
ened (if pain changed from no pain to mild, moderate, or
severe, if pain changed from mild to moderate or severe,
or if pain changed from moderate to severe); (2) no
change (if the pain level category did not change from
baseline to follow-up); (3) improved (if pain changed
from severe to moderate, severe to mild, or moderate to
mild); and (4) resolved (if pain changed from moderate,
mild, or severe to no pain). Changes in depression levels
were categorized in the same manner, ie, worsened, no
change, improved, and resolved.
All data were converted from the SAS database to
STATA for statistical analysis (STATA Corporation:
analysis to identify outliers, normal or non-normal
distributions, data entry errors, missing data, and bio-
logic plausibility of unusual values. Continuous variable
distributionsweretested fornormality usingtheShapiro-
Wilk statistic. Continuous variables with skewed distri-
butions were summarized using medians and ranges and
categorical data were described using frequencies.
Bivariate analyses were performed to determine as-
sociations between participant characteristics and treat-
ment group. The Wilcoxon rank sum test was used for
non-normal continuous variables, Pearson’s c2was used
for comparison of categorical variables, and Fisher
exact test for comparison of categorical variables with
low frequencies. The comparison of before and after
pain levels used 1-sample paired t tests for within group
analysis and 2-sample t tests for comparisons between 2
groups. Bivariate analyses were used to calculate odds
ratios (OR) and 95% confidence intervals (95%CI) for
the association between treatment group and improve-
ment in pain and depression levels.
Multivariable analyses were performed to describe
the association of having surgical treatment with
improvement of pain. To identify potential effect mod-
ifiers we first conducted bivariate comparisons between
all patient characteristics and the exposure, ie, having
surgical treatment. Effect modifiers were considered if
they had a Mantel-Haenszel test for homogeneity of
P ! .1. Next, we conducted bivariate analysis to iden-
tify any potential relationship between patient charac-
teristics and improved post-treatment pain levels. Any
patient characteristic that was associated (P ! .1)
with having surgical treatment or with having the out-
come was evaluated in the final multivariable models as
a possible confounder of the relationship between treat-
ment group and improvement in pain. Confounders
evaluated in final models included age, history of sexual
abuse, history of physical abuse, current depression
(moderate or severe), current sexual dysfunction, previ-
ous treatments tried (medical, surgical or both), num-
ber of pain sites, and duration of pain. We explored
Poisson regression and GEE models where improve-
ment in pain level was defined as a decrease in the 1-
year MPQ score by 1 point, or 4 points, or 30%
from baseline. Later we also generated unconditional
logistic regression models using improvement in pain
level as the dependent variable. We started with a full
model of all variables identified as potential con-
founders of the surgical treatment-improved pain asso-
ciation and used backward elimination to remove
1 variable at a time. A 10% change in the point esti-
mate and likelihood ratios was used to determine
whether a confounder should be retained. Because
our conclusions were generally the same for all models
Lamvu et al 593
Characteristics of women evaluated for chronic pelvic pain that completed evaluation 1 year after medical or surgical
P value* (n = 370) % (95% CI)(n = 181) % (95% CI) (n = 189)% (95% CI)
Mean age (y)
(SD, Range, 95% CI)
High school graduate
Mean duration of pain (y)
(SD, Range, 95% CI)
Duration of pain (y)
Mean number of pain sites
Number of pain sites
Had previous surgical
Sexual abuse, adult
Physical abuse, adult
Most common diagnosis
Pelvic floor tension
33.3 (10.0, 14-70, 32.3-34.3)34.8 (11.6, 15-70, 33.1-36.5)31.9 (8.0, 14-59, 30.8-33.0).003
7 4 (2-8)
6 3 (1-7)
4.9 (6.3, 0.1-34, 4.0-5.8)
4.4 (4.4, 0.1-20, 3.8-5.0)
4.6 (5.4, 0.1-34, 4.0-5.2).422
3.8 (1.9, 1-8, 3.6-4.0)
3.6 (1.8, 1-8, 3.4-3.9)
3.9 (1.9, 1-8, 3.6-4.2)
13336 (31-41) 6838 (31-45) 6534 (28-42).732
GED, General equivalency diploma.
* P values represent comparisons between the medical and surgical groups.
594Lamvu et al
explored, the results presented here are based on a final
multivariable regression model where the pain outcome
was a bivariate variable that categorized all patients
pain levels into improved (if the pain resolved or im-
proved) or not improved (if the pain worsened or re-
Of the 987 women seen in our CPP clinic between 1993
and 2000, records were successfully abstracted in 970
(98%) cases. The data presented here are obtained from
370 (38%) women whose records were complete and
who returned the questionnaires at 1 year. Baseline
characteristics of questionnaire respondents (n = 370)
and nonrespondents (n = 600), were statistically
similar in age, racial distribution, marital status, annual
income, education, duration of pain, number of pain
sites, history of child or adult sexual abuse, history of
adult physical abuse, depression, sexual dysfunction,
and previous treatments (including number of previous
In the 370 women who were evaluated at baseline and
1 year, mean age was 33 years (range 14-70), 84% were
Caucasian, 24% had never been married, 33% had an
annual income less than $20,000, and 45% had less than
a high-school degree (Table I).
For the entire study cohort, the mean baseline MPQ
score was 30.4 (SD 14.7, range 3-84). Thirty percent of
patients had pain for 4 years or more, 33% had at least 5
painful sites, and 93% had previous surgery for their
pain before referral to the UNC CPP clinic. Seventy-six
percent of women reported sexual dysfunction, defined
as painful intercourse, decreased frequency, or decreased
pleasure resulting from pain.
Twenty-two percent of women had moderate or severe
depression defined by the BDI. Of the women we could
were associated with higher MPQ scores at baseline.
After evaluation in the CPP clinic, the most common
diagnosis made included: irritable bowel syndrome, ad-
hesions, pelvic floor musculoskeletal disorders, and en-
181) of the patients received medical treatment (pharma-
cotherapy, physical therapy, cognitive-behavioral ther-
apy, or combinations of the 3) and 51% (n = 189)
received surgical treatment (54 hysterectomies and 135
other procedures without hysterectomy, eg, diagnostic
laparoscopy, adhesiolysis, resection of endometriosis,
oophrectomy). After 1 year, the mean MPQ score for
the entire cohort decreased from 30.4 to 23.3 (SD 17.3,
range 0-75), P ! .0001.
The medical and surgical treatment groups were
comparable in marital status, income, and education
(Table I). However, women in the medical group were
older than those in the surgical group (38.4 yrs vs.
31.9 yrs, P = .003). Dyspareunia, adhesions, endometri-
osis, and pelvic floor tension were more commonly diag-
nosed during clinical evaluation in those who had
surgery at UNC (Table I).
At baseline, pain intensity and duration was similar
in both groups (Table I). However, the surgical group
had a higher frequency of women with 5 or more pain
sites (38% vs 28%, P = .038) when compared with
the medical group (Table I). The prevalence of sexual
abuse, physical abuse, depression, and sexual dysfunc-
tion was similar in both groups. At 1 year, for 48% of
patients (medical and surgical groups) the severity of
depression did not change, and in 21% the depression
worsened (Table II). For 31% the depression improved
or resolved predominately in those with mild depression
at initial evaluation (Table II).
McGill pain scores were significantly lower after
1 year of treatment within each treatment group (med-
ical group P ! .000 and surgical group P ! .000), but
improvement was similar in both groups (P = .165).
Pain levels were moderate or severe in 49% of patients
in both the medical and surgical groups at baseline and
decreased to less than 34% in both groups after 1 year
(Table III). Nineteen percent of the medical group and
16% of the surgical group reported worsening pain
All Medical Surgical
P value* (n = 370) % (95% CI)(n = 181)% (95% CI) (n = 189) % (95% CI)
Low back pain
GED, General equivalency diploma.
* P values represent comparisons between the medical and surgical groups.
Lamvu et al 595