American Journal of Epidemiology Vol. 165, No.

10
Copyright ª 2007 by the Johns Hopkins Bloomberg School of Public Health DOI: 10.1093/aje/kwm011
All rights reserved; printed in U.S.A. Advance Access publication March 20, 2007

Original Contribution

Risk Factors for Inguinal Hernia among Adults in the US Population

Constance E. Ruhl1 and James E. Everhart2

1
Social & Scientific Systems, Inc., Silver Spring, MD.
2
National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD.

Received for publication July 7, 2006; accepted for publication November 14, 2006.

Downloaded from http://aje.oxfordjournals.org/ by guest on October 23, 2015

The authors examined risk factors for incident inguinal hernia among US adults (5,316 men and 8,136 women)
participating in the First National Health and Nutrition Examination Survey (1971–1975) who were followed through
1992–1993 for a hospital (International Classification of Diseases, Ninth Revision, Clinical Modification, code 550)
or physician diagnosis of inguinal hernia. Ninety-six percent of the baseline cohort was recontacted, with a median
follow-up of 18.2 years (range, 0.02–22.1 years). Because the cumulative incidence of inguinal hernia was higher
among men (13.9%) than among women (2.1%), more detailed analyses were conducted in men. Among men in
multivariate analysis, a higher incidence (p < 0.05) of inguinal hernia was associated with an age of 40–59 years
(hazard ratio (HR) ¼ 2.2, 95% confidence interval (CI): 1.7, 2.8), an age of 60–74 years (HR ¼ 2.8, 95% CI: 2.2,
3.6), and hiatal hernia (HR ¼ 1.8, 95% CI: 1.2, 2.7), while Black race (HR ¼ 0.58, 95% CI: 0.42, 0.79), being
overweight (HR ¼ 0.79, 95% CI: 0.66, 0.95), and obesity (HR ¼ 0.51, 95% CI: 0.36, 0.71) were associated with
a lower incidence. Among women, older age, rural residence, greater height, chronic cough, and umbilical hernia
were associated with inguinal hernia. In the United States, inguinal hernias are common among men, especially
with aging. The lower risk among heavier men was unexpected and bears further study.

adult; hernia, inguinal; prospective studies; risk factors; United States

Abbreviations: CI, confidence interval; HR, hazard ratio; NHANES I, First National Health and Nutrition Examination Survey.

Inguinal hernias are the most common form of abdominal
wall hernias. The incidence of inguinal hernia is unknown,
but about 500,000 cases come to medical attention each year
(1). No recent data on the prevalence of inguinal hernia in the
United States based on physical examination are available.
In international and US surveys conducted 20 or more years
ago, the prevalence of non-surgically-treated inguinal hernia
among men was 5–7 percent, and a similar number of men
had a history of hernia repair (1). Inguinal hernias are much
more common among men than among women. They may
also be more common among Whites and older adults (1).
The health effects of inguinal hernia are considerable. In
1989–1990, there were an estimated 1.65 million annual
first-listed ambulatory care visits for inguinal hernia in the
United States (1). Today most herniorrhaphies are per-
formed as an outpatient procedure. Before this trend began,
there were over 600,000 overnight hospital stays per year for
inguinal hernia in the United States (1). Furthermore, it has
been estimated that in the United States, hernias have re-
sulted in significant limitation of activity for approximately
400,000 persons, and the number of days of work lost is
higher than for any other chronic digestive condition (1).
Complications of inguinal hernia include incarceration,
bowel obstruction, and bowel strangulation (which is poten-
tially fatal), with the greatest risk being found among older
persons. Although risk of death is small, hernia was listed
as the underlying cause of death for 1,595 US deaths in
2002 (2).
Despite the common occurrence and clinical significance
of inguinal hernia, only a few studies have investigated risk

Correspondence to Dr. Constance E. Ruhl, Social & Scientific Systems, Inc., 8757 Georgia Avenue, 12th floor, Silver Spring, MD 20910
(e-mail: cruhl@s-3.com).

1154 Am J Epidemiol 2007;165:1154–1161


FIGURE 1. Cumulative probability of inguinal hernia by sex among adults in the United States, 1971–1993.
factors for inguinal hernia (3–7). Two studies found in-
creased risk with strenuous exertion (4, 5). Interestingly,
being overweight was associated with lower risk in two
studies (3, 6). Associations with inguinal hernia were found
in individual studies for varicose veins (3), history of hem-
orrhoids (3), smoking (8), and hiatal hernia (7). To our
knowledge, risk factors for inguinal hernia have not been
evaluated in a prospective cohort study or in any US popu-
lation. Therefore, we examined the relation between poten-
tial risk factors and incident inguinal hernia in the First
National Health and Nutrition Examination Survey
(NHANES I) and its follow-up study, a large, population-
based US prospective study.
MATERIALS AND METHODS
NHANES I
Conducted between 1971 and 1975, NHANES I included
interview, examination, and laboratory data collected from
a national probability sample of the civilian, noninstitution-
alized US population (9, 10). Of 20,729 sampled persons,
14,407 (70 percent) underwent a medical examination (11).
At the NHANES I interview, participants were asked about
nonrecreational activity (inactive, moderately active, very
active) and recreational activity (little or no exercise, mod-
erate exercise, much exercise); constipation and frequency
of bowel movements; and whether they had ever been told
by a doctor that they had a hiatal hernia of the diaphragm,
a chronic cough, or chronic bronchitis or emphysema. Par-
ticipants were also asked about their frequency and quantity
of beer, wine, and liquor consumption; alcohol use (number
of drinks per day) was summarized as none, <1, 1–2, or >2.
Women were asked whether their menstrual periods had
stopped entirely. Weight (kg), standing height (cm), and
sitting height (cm) were measured. Body mass index (weight
Am J Epidemiol 2007;165:1154–1161
Risk Factors for Inguinal Hernia in US Adults 1155
Downloaded from http://aje.oxfordjournals.org/ by guest on October 23, 2015
(kg)/(height (m)2) was calculated, and participants were cat-
egorized as normal weight (<25), overweight (25–<30), or
obese (
30). The presence of an umbilical hernia was noted
upon physical examination. Demographic variables were
age (25–39, 40–59, or 60–74 years), sex, race/ethnicity
(White, Black, other), education (less than high school grad-
uation, high school graduation or more), and urbanicity (ur-
ban or rural residence). Information on smoking (never,
former, current) obtained in a subgroup was supplemented
by responses to questions on lifetime smoking history posed
to all participants in the 1982–1984 interview (12). A ran-
domly chosen subgroup of 79 percent of participants were
asked about their minimum and maximum adult weights.
Weight loss from the maximum adult weight before baseline
and weight gain from the minimum adult weight before
baseline were calculated from reported weights and mea-
sured baseline weight.
NHANES I Epidemiologic Follow-up Study
The NHANES I Epidemiologic Follow-up Study was
a longitudinal study of the 14,407 NHANES I participants
aged 25–74 years who had been medically examined. It was
conducted in four waves: 1982–1984, 1986, 1987, and
1992–1993 (11, 13–15). Ninety-six percent of the baseline
cohort was recontacted. Participants were interviewed and
hospital and nursing home records were collected at each
wave. Records were obtained for all overnight medical fa-
cility stays occurring since the participant’s NHANES I ex-
amination. Hospital discharge and nursing home admission
diagnoses were recoded by trained medical coders using the
International Classification of Diseases, Ninth Revision,
Clinical Modification (16). We defined a case of inguinal
hernia as any facility stay with an International Classifica-
tion of Diseases, Ninth Revision, Clinical Modification, di-
agnosis that had a three-, four-, or five-digit code for which
the first three digits were 550, inguinal hernia. Because
1156 Ruhl and Everhart

TABLE 1. Cumulative probability (unadjusted) of inguinal hernia over approximately 20 years and age-
adjusted hazard ratio for inguinal hernia among men in the United States (n ¼ 5,316), 1971–1993

Cumulative Age-
No. with 95%
No. of incidence of adjusted p for
Characteristic inguinal confidence
participants inguinal hazard trendy
hernia interval
hernia (%) ratio*
Age (years) <0.001
25–39 1,417 92 7.3 1.0
40–59 1,818 209 14.8 2.1 1.6, 2.7
60–74 2,081 199 22.8 2.7 2.1, 3.5
Race/ethnicity NAz
White 4,481 457 15.1 1.0
Black 757 42 8.4 0.56 0.40, 0.76
Other 78 1 1.8 0.12 0.017, 0.85
Education (years) NA
<12 2,625 236 15.6 1.0

12 2,644 261 12.8 1.1 0.93, 1.4

Downloaded from http://aje.oxfordjournals.org/ by guest on October 23, 2015

Height (cm)§ 0.20
<171 1,755 155 13.2 1.0
171–<177 1,776 175 14.5 1.2 0.95, 1.5

177 1,783 170 14.5 1.2 0.92, 1.5
Weight (kg)§ 0.022
<71.2 1,771 162 13.5 1.0
71.2–<82.3 1,772 207 16.9 1.2 0.96, 1.5

82.3 1,772 131 11.6 0.76 0.60, 0.95
Body mass index{ <0.001
<25 2,380 248 15.6 1.0
25–<30 2,246 212 14.0 0.82 0.68, 0.98

30 688 40 8.3 0.53 0.38, 0.74
Nonrecreational physical activity 0.42
Low 603 36 11.1 1.0
Moderate 2,366 231 14.2 1.3 0.92, 1.9
High 2,342 233 14.2 1.3 0.90, 1.8
Recreational physical activity 0.46
Low 1,998 172 13.3 1.0
Moderate 2,082 206 15.6 1.1 0.89, 1.3
High 1,230 122 12.6 1.1 0.86, 1.4
Maximum adult weight (kg)§,# 0.031
<77 1,171 113 14.0 1.0
77–<90 1,538 174 16.3 1.2 0.93, 1.5

90 1,372 96 11.5 0.74 0.56, 0.97
Minimum adult weight (kg)§,# 0.25
<61 1,180 118 17.8 1.0
61–<70 1,465 138 13.6 0.91 0.71, 1.2

70 1,327 118 11.8 0.86 0.66, 1.1
Table continues

outpatient hernia repairs became increasingly common over and if so, in what year they were first told of the diagnosis.
the 20 years of follow-up, participants were asked at the Of 620 inguinal hernia cases, 373 were first-facility-stay
1992 interview if they had ever been told by a doctor that diagnoses (99 percent were hospitalizations and 1 percent
they had a hernia or rupture of the groin (inguinal hernia), were nursing home admissions) and 247 were physician

Am J Epidemiol 2007;165:1154–1161
 Risk Factors for Inguinal Hernia in US Adults 1157

TABLE 1. Continued

Cumulative Age-
No. with 95%
No. of incidence of adjusted p for
Characteristic inguinal confidence
participants inguinal hazard trendy
hernia interval
hernia (%) ratio*

Weight loss from maximum lifetime 0.46

weight to baseline (kg)§
<4.1 1,360 126 11.8 1.0
4.1–<9.1 1,360 140 16.1 1.1 0.90, 1.5

9.1 1,361 117 14.1 1.1 0.85, 1.4
Weight gain from minimum weight since 0.37
age 18 years to baseline (kg)§
<5.1 1,324 123 14.3 1.0
5.1–<13.7 1,323 130 13.6 0.97 0.76, 1.2

13.7 1,325 121 14.4 0.89 0.69, 1.1
Alcohol drinking (no. of drinks/day) 0.90
0 (nondrinker) 1,636 152 14.1 1.0

Downloaded from http://aje.oxfordjournals.org/ by guest on October 23, 2015

>0–<1 2,200 217 14.9 1.1 0.85, 1.3
1–2 820 78 12.3 1.0 0.79, 1.4
>2 634 51 12.0 0.96 0.69, 1.3
Cigarette smoking NA
Never smoker 1,340 139 13.4 1.0
Former smoker 1,380 173 18.0 1.2 0.93, 1.5
Current smoker 1,974 150 12.1 0.82 0.65, 1.0
Hiatal hernia NA
No 5,169 475 13.6 1.0
Yes 136 23 26.9 1.8 1.2, 2.8
Chronic cough NA
No 5,103 487 14.1 1.0
Yes 208 13 8.6 0.74 0.42, 1.3
Chronic bronchitis or emphysema NA
No 4,965 475 13.9 1.0
Yes 326 23 15.2 0.83 0.55, 1.3
Constipation NA
No 4,700 448 14.0 1.0
Yes 602 52 14.3 0.93 0.70, 1.2
Bowel movement frequency (per day) NA
1 3,810 364 14.4 1.0
<1 614 53 12.4 0.99 0.74, 1.3
>1 859 81 13.6 0.98 0.77, 1.3

* Estimated using Cox proportional hazards regression analysis, adjusted for 10-year age group (age categories
were unadjusted).
y Calculated by including the factor in the analysis as an ordinal variable with three or more levels.
z NA, not applicable.
§ Cutpoints were tertiles (thirds).
{ Weight (kg)/height (m)2.
# Data were collected in a random subgroup.

diagnoses reported at the 1992 interview among persons date of first physician diagnosis among persons without
without a preceding facility-stay diagnosis. The date of di- a documented facility-stay diagnosis.
agnosis was considered to be the admission date of the first Excluded from the current analysis were 546 survey par-
facility stay with an inguinal hernia diagnosis or the reported ticipants who could not be traced, 288 who in 1992 reported

Am J Epidemiol 2007;165:1154–1161
1158 Ruhl and Everhart
FIGURE 2. Cumulative probability of inguinal hernia by age among men in the United States, 1971–1993.
a doctor-diagnosed inguinal hernia first diagnosed in a year
prior to the year of their NHANES I examination, and 121
who in 1992 reported a doctor-diagnosed inguinal hernia
with the year of first diagnosis unknown. Remaining for
analysis were 13,452 persons (5,316 men and 8,136 women)
with a median follow-up time of 18.2 years (range, 0.02–
22.1 years).
Statistical analysis
For each level of potential risk factors, we calculated the
cumulative percentage of participants with inguinal hernia
during the 20 years of follow-up. We estimated curves for
percentage with inguinal hernia (reciprocal of the survival
curve) by categories of possible risk factors using Kaplan-
Meier analysis. We then calculated hazard ratio estimates
using Cox proportional hazards regression models in SAS
(PROC PHREG and SAS OnlineDoc, version 8; SAS In-
stitute, Inc., Cary, North Carolina) to take into consideration
varying lengths of follow-up. Time at risk was calculated
from the date of the NHANES I examination to the date of
inguinal hernia diagnosis for cases or the date of last contact
or death for noncases. All factors met the proportional haz-
ards assumption of a relatively constant risk ratio through
examination of -log (-log) plots of survival by duration of
follow-up (17). For each possible risk factor, we conducted
analysis adjusting for age (in 10-year groups) as an ordinal
variable. The relation of incident inguinal hernia with po-
tential risk factors was examined further in multivariate pro-
portional hazards analysis while controlling for effects of
multiple factors. Factors that were related to inguinal hernia
in age-adjusted analysis (p < 0.10) were included in multi-
variate models. Multivariate analyses excluded persons with
missing values for any risk factor included in the model. The
trend in the relative risk of inguinal hernia across categories
of risk factors was computed by including each factor in
analyses as an ordinal variable with multiple levels. Because
cases were based on physician diagnoses, either identified
Downloaded from http://aje.oxfordjournals.org/ by guest on October 23, 2015
from hospital records or reported at the 1992 follow-up, and
persons who are hospitalized or visit a physician are a se-
lected group, we repeated the analyses in the population
subgroup with one or more facility stays.
RESULTS
The cumulative incidence of hospitalization with inguinal
hernia was 6.3 percent at 20 years and was much higher
among men (13.9 percent; 500 cases) than among women
(2.1 percent; 120 cases) (figure 1). Adjusted for age, the
inguinal hernia hazard ratio for men relative to women
was 7.5 (95 percent confidence interval (CI): 6.2, 9.2). Con-
sequently, more detailed analyses were conducted in men.
The 20-year cumulative incidence of inguinal hernia among
men increased with baseline age: 7.3 percent at age 24–39
years, 14.8 percent at age 40–59 years, and 22.8 percent at
age 60–74 years (table 1, figure 2). White men had almost
twice the 20-year cumulative incidence (15.1 percent) of
Black men (8.4 percent). The relation of inguinal hernia with
possible risk factors was examined in age-adjusted analysis.
A lower incidence of inguinal hernia was associated with
higher levels of some weight-related factors, including body
mass index and maximum lifetime weight (table 1, figure 3).
Men reporting a doctor-diagnosed hiatal hernia had a higher
incidence of inguinal hernia. Current smokers had a border-
line lower risk of inguinal hernia. Factors that were unrelated
to inguinal hernia included nonrecreational and recreational
physical activity, constipation or bowel movement fre-
quency, chronic cough, chronic obstructive pulmonary dis-
ease, alcohol intake, minimum adult weight, weight loss or
weight gain, height, and education.
In multivariate analysis among men, greater age and
a doctor-diagnosed hiatal hernia remained associated with
a higher incidence of inguinal hernia, while Black race and
overweight and obesity remained associated with a lower
incidence (table 2). Middle-aged men had over twice the
Am J Epidemiol 2007;165:1154–1161
 Risk Factors for Inguinal Hernia in US Adults 1159

FIGURE 3. Cumulative probability of inguinal hernia by body mass index (weight (kg)/height (m)2) among men in the United States, 1971–1993.

Downloaded from http://aje.oxfordjournals.org/ by guest on October 23, 2015

incidence of younger men, and the risk increased to almost facility-stay diagnoses, 67 percent of the number without
three times among older men. The presence of a hiatal her- this restriction. Results were similar in direction to those
nia almost doubled the risk. Black men had slightly more for the entire population, though they no longer reached
than half the incidence of White men. Obese men had half statistical significance for Black race or doctor-diagnosed
the risk of normal-weight men, while incidence was inter- hiatal hernia. In multivariate-adjusted analysis, a higher in-
mediate among the overweight. Recoding body mass index cidence of inguinal hernia was found with middle age (haz-
as a continuous variable resulted in a 4.7 percent lower risk ard ratio (HR) ¼ 1.7, 95 percent CI: 1.2, 2.3) and older age
of inguinal hernia for each unit increase in body mass index (HR ¼ 2.3, 95 percent CI: 1.6, 3.2), while a lower incidence
(p < 0.001). Other weight-related factors were not associ- was found with overweight (HR ¼ 0.76, 95 percent CI: 0.61,
ated with inguinal hernia independently of body mass index. 0.95) and obesity (HR ¼ 0.50, 95 percent CI: 0.34, 0.75).
When analysis was limited to men with one or more med- Associations with Black race (HR ¼ 0.74, 95 percent CI:
ical facility stays (n ¼ 3,452), there were 337 inguinal hernia 0.52, 1.1) and hiatal hernia (HR ¼ 1.5, 95 percent CI: 0.92,
2.5) were of borderline statistical significance.
Among women, older age was associated with a greater
TABLE 2. Multivariate-adjusted hazard ratio for inguinal
incidence of inguinal hernia. Other factors associated with
hernia among men in the United States (n ¼ 5,303), 1971–1993 a higher inguinal hernia incidence in age-adjusted analyses
were rural residence, greater height, chronic cough, umbil-
Variable Hazard ratio*
95% confidence ical hernia, greater sitting height, and postmenopausal sta-
interval
tus. Sitting height and postmenopausal status were highly
Age (years) correlated with standing height and age, respectively, and
24–39 1.0 therefore were not included in multivariate-adjusted analy-
40–59 2.2 1.7, 2.8 ses. Middle and older age, rural residence, height in the
60–74 2.8 2.2, 3.6
upper two thirds, chronic cough, and umbilical hernia re-
mained independently associated with higher incidence of
Race/ethnicity
inguinal hernia in multivariate-adjusted analysis (table 3).
White 1.0 Among women, multivariate-adjusted hazard ratios for
Black 0.58 0.42, 0.79 Black race (HR ¼ 0.68, 95 percent CI: 0.37, 1.3) and hiatal
Body mass indexy hernia (HR ¼ 1.7, 95 percent CI: 0.74, 4.0) were similar to
<25 1.0 those among men but did not reach statistical significance.
Overweight (HR ¼ 0.99, 95 percent CI: 0.64, 1.5) and obe-
25–<30 0.79 0.66, 0.95
sity (HR ¼ 1.1, 95 percent CI: 0.71, 1.8) were unrelated to

30 0.51 0.36, 0.71 inguinal hernia among women.
Hiatal hernia
No 1.0
Yes 1.8 1.2, 2.7 DISCUSSION

* Estimated using Cox proportional hazards regression analysis. Possibly the clearest message of this study is how com-
The model included all variables listed in the table. monly inguinal hernia occurs. Many health-care providers
y Weight (kg)/height (m)2. may consider inguinal hernia to be largely limited to male

Am J Epidemiol 2007;165:1154–1161
1160 Ruhl and Everhart

TABLE 3. Multivariate-adjusted hazard ratio for inguinal a protective effect of greater adiposity exists in that among
hernia among women in the United States (n ¼ 8,104), heavier men, abdominal wall musculature may be strength-
1971–1993 ened by carrying excess fat, providing a stronger barrier
95% confidence against herniation. If this is true, one might expect a central
Variable Hazard ratio*
interval fat distribution, in particular, to be protective. However, we
Age (years) did not have information available on waist circumference
or other measures of abdominal fat. More research is needed
24–39 1.0
to determine whether the association of inguinal hernia with
40–59 1.9 1.3, 3.0 overweight and obesity is real. The effect of body fat dis-
60–74 2.0 1.2, 3.3 tribution needs to be investigated as well.
Urbanicity The incidence of inguinal hernia was much higher in men
Urban residence 1.0 than in women, as has been previously shown (2). Factors
Rural residence 1.8 1.3, 2.6
that were independently associated with a higher incidence
of inguinal hernia among women were middle or older age,
Height (cm)y
rural residence, height in the upper two thirds, chronic
<158 1.0 cough, and umbilical hernia.
158–<164 2.0 1.2, 3.2 Among men, we found an incidence of inguinal hernia

164 1.9 1.2, 3.2 among Blacks that was less than 60 percent of that of
Whites. This lower rate of hernia among Black men could

Downloaded from http://aje.oxfordjournals.org/ by guest on October 23, 2015

Chronic cough
No 1.0
Yes 2.8 1.3, 6.0
Umbilical hernia
No 1.0
Yes 3.2 1.2, 8.7
* Estimated using Cox proportional hazards regression analysis.
The model included all variables listed in the table.
y Cutpoints were tertiles (thirds).
infants with an anatomical defect in the inguinal canal (18).
In contrast, we have documented that inguinal hernia occurs
frequently among adults, such that well over one quarter of
adult men in the United States would be expected to have a
medically recognized inguinal hernia. In fact, the incidence
of inguinal hernia increased noticeably with age among men.
An unexpected finding was the lower incidence of ingui-
nal hernia among overweight and obese men as compared
with normal-weight men. The risk among overweight men
was 80 percent of that of normal-weight men, and the risk
among obese men was only 50 percent of that of normal-
weight men. It is possible that our findings resulted from
ascertainment bias due to greater difficulty in diagnosing
hernia among heavier persons. On the other hand, over-
weight and obese persons have more comorbidity requiring
medical care and have a greater opportunity for a hernia
diagnosis, which actually could have led to underestimation
of the strength of a protective effect in our analysis. Among
men with at least one facility stay, the lower risks of hernia
among the overweight and obese were unchanged, while the
strength of association of age, race, and hiatal hernia were
diminished. Thus, receipt of medical care does not explain
the lower risk of inguinal hernia among the overweight and
obese.
A lower risk of inguinal hernia with overweight and obe-
sity was also suggested in a community survey of men in
Israel (3) and in a hospital-based case-control study of
women in the Netherlands (6). A plausible explanation for
have resulted from a lower rate of utilization of medical
care, with less opportunity for a hernia diagnosis. Among
men with at least one facility stay, Blacks had a lower risk of
hernia, although this result no longer reached statistical sig-
nificance. Racial differences in body mass index could also
have contributed, but the lower incidence in Blacks re-
mained in multivariate-adjusted analysis.
The final factor that we found to be associated with
a greater incidence of inguinal hernia was doctor-diagnosed
hiatal hernia, which almost doubled the risk. Some men who
reported a doctor-diagnosed hiatal hernia at baseline may
have confused it with a history of inguinal hernia. However,
an association between the two types of hernias was also
seen in an Italian case-control study of endoscopy-diagnosed
hiatal hernia (7). In that report, the risk of inguinal hernia
upon physical examination was increased 2.5-fold among
persons with hiatal hernia, though the older age of the hiatal
hernia patients may have confounded the relation. An asso-
ciation between these two forms of hernia could be due to
a common mechanism of increased intraabdominal pressure.
Increased intraabdominal pressure has long been sus-
pected in the pathogenesis of inguinal hernia, though with
little quantitative evidence. We investigated but did not find
an association with additional factors that might exert an
effect through such a mechanism, including physical activ-
ity, constipation, chronic cough, and chronic obstructive
pulmonary disease. Our physical activity measure was lim-
ited to two interview questions, each with three possible
subjective responses. An increased risk of inguinal hernia
with greater physical exertion was found in two Spanish
hospital-based case-control studies investigating occupa-
tional activity (4) or both work and recreational activity
(5), while greater current sports activity was found to de-
crease the risk among Dutch women (6). There was no re-
lation with work-related physical activity among Israeli men
(2). Other factors that might increase intraabdominal pres-
sure were not associated with inguinal hernia in previous
studies, with the exception of an increased risk with obsti-
pation in the Dutch study (3, 5, 6).
Structural weakness of the supporting tissue is another
potential mechanism in the pathogenesis of inguinal hernia.

Am J Epidemiol 2007;165:1154–1161

A defect in collagen synthesis by fibroblasts has been sug-
gested as a cause of inguinal hernia (19, 20). Smoking,
which may adversely affect connective tissue metabolism,
has been proposed as a risk factor for inguinal hernia (21)
and was associated with hernia recurrence among smokers
in one study (8). We did not find an increased incidence of
inguinal hernia among current or former smokers. However,
smoking may have a greater adverse effect on connective
tissue healing than on intact connective tissue. Smoking was
also unrelated to a first hernia diagnosis in previous reports
(5, 6). Likewise, no relation of inguinal hernia with alcohol
intake was seen in our study or in a previous one (5).
Our study had limitations. Because follow-up occurred
between 10 and 20 years after the baseline examination, it
is possible that some participants may have developed an
inguinal hernia and died of complications, or may have been
lost to follow-up for other reasons related to the hernia, in the
intervening period. However, the follow-up rate was high
(96 percent of persons in the baseline cohort were recon-
tacted) and proxies were interviewed in the case of deceased
participants, so this should have minimally biased the re-
sults. Secondly, because the case definition was based on
hospital diagnoses and self-reported physician diagnoses,
case ascertainment may have been incomplete. Confirmation
of diagnoses by physical examination and chart review was
not possible. Finally, although rates of participation and
follow-up were high, participation and follow-up were in-
complete, which decreases generalizability. Despite these
limitations, our study contributes prospective data from a
US population to the limited body of literature on risk factors
for inguinal hernia.
In conclusion, in the US population, inguinal hernia is
a common condition among men that increases substantially
with aging. The lower risk among overweight and obese
men was an unexpected finding that bears further study.
ACKNOWLEDGMENTS
This work was supported by a contract (N01-DK-1-
2478) with the National Institute of Diabetes and Digestive
and Kidney Diseases.
The authors thank Danita Byrd-Holt for computer pro-
gramming assistance.
Conflict of interest: none declared.
REFERENCES
1. Everhart JE. Abdominal wall hernia. In: Everhart JE, ed. Di-
gestive diseases in the United States: epidemiology and im-
pact. Bethesda, MD: National Institute of Diabetes and
Digestive and Kidney Diseases, 1994:471–507.
2. Kochanek KD, Murphy SL, Anderson RN, et al. Deaths: final
data for 2002. Hyattsville, MD: National Center for Health
Statistics, 2004. (National vital statistics reports, vol 53, no. 5).
3. Abramson JH, Gofin J, Hopp C, et al. The epidemiology of
inguinal hernia. A survey in western Jerusalem. J Epidemiol
Community Health 1978;32:59–67.
Am J Epidemiol 2007;165:1154–1161
Risk Factors for Inguinal Hernia in US Adults 1161
4. Flich J, Alfonso JL, Delgado F, et al. Inguinal hernia and
certain risk factors. Eur J Epidemiol 1992;8:277–82.
5. Carbonell JF, Sanchez JL, Peris RT, et al. Risk factors asso-
ciated with inguinal hernias: a case control study. Eur J Surg
1993;159:481–6.
6. Liem MS, van der Graaf Y, Zwart RC, et al. Risk factors for
inguinal hernia in women: a case-control study. The Coala
Trial Group. Am J Epidemiol 1997;146:721–6.
7. De Luca L, Di Giorgio P, Signoriello G, et al. Relationship
between hiatal hernia and inguinal hernia. Dig Dis Sci
2004;49:243–7.
8. Sorensen LT, Friis E, Jorgensen T, et al. Smoking is a risk
factor for recurrence of groin hernia. World J Surg 2002;
26:397–400.
9. Miller HW. Plan and operation of the Health and Nutrition
Examination Survey, United States 1971–1973. Part A—
development, plan, and operation. Hyattsville, MD: National
Center for Health Statistics, 1973. (Vital and health statistics,
series 1, no. 10a) (DHEW publication no. (PHS) 76-1310).
10. Engel A, Murphy RS, Maurer K, et al. Plan and operation of
Downloaded from http://aje.oxfordjournals.org/ by guest on October 23, 2015
the NHANES I Augmentation Survey of adults 25–74 years,
United States 1974–1975. Hyattsville, MD: National Center
for Health Statistics, 1978. (Vital and health statistics, series 1,
no. 14) (DHEW publication no. (PHS) 78-1314).
11. Cohen BB, Barbano HE, Cox CS, et al. Plan and operation of
the NHANES I Epidemiologic Followup Study, 1982–84.
Hyattsville, MD: National Center for Health Statistics, 1987.
(Vital and health statistics, series 1, no. 22) (DHEW publica-
tion no. (PHS) 87-1324).
12. McLaughlin JK, Dietz MS, Mehl ES, et al. Reliability of
surrogate information on cigarette smoking by type of infor-
mant. Am J Epidemiol 1987;126:144–6.
13. Finucane FF, Freid VM, Madans JH, et al. Plan and operation
of the NHANES I Epidemiologic Followup Study, 1986. Hy-
attsville, MD: National Center for Health Statistics, 1987.
(Vital and health statistics, series 1, no. 25) (DHEW publica-
tion no. (PHS) 90-1307).
14. Cox CS, Rothwell ST, Madans JH, et al. Plan and operation of
the NHANES I Epidemiologic Followup Study, 1987. Hy-
attsville, MD: National Center for Health Statistics, 1992.
(Vital and health statistics, series 1, no. 27) (DHEW publica-
tion no. (PHS) 92-1303).
15. Cox CS, Mussolino ME, Rothwell ST, et al. Plan and operation
of the NHANES I Epidemiologic Followup Study, 1992. Hy-
attsville, MD: National Center for Health Statistics, 1997.
(Vital and health statistics, series 1, no. 35) (DHEW publica-
tion no. (PHS) 98-1311).
16. US Department of Health and Human Services. International
classification of diseases, ninth revision, clinical modification.
4th ed. Vol 1. Tabular list of diseases. Washington, DC: US
GPO, 1991. (DHHS publication no. (PHS) 91-1260).
17. Ingram DD, Makuc DM. Statistical issues in analyzing the
NHANES I Epidemiologic Followup Study. Hyattsville, MD:
National Center for Health Statistics, 1994. (Vital and health
statistics, series 2, no. 121) (DHHS publication no. (PHS)
94-1395).
18. Graf JL, Caty MG, Martin DJ, et al. Pediatric hernias. Semin
Ultrasound CT MR 2002;23:197–200.
19. Wagh PV, Read RC. Defective collagen synthesis in inguinal
herniation. Am J Surg 1972;124:819–22.
20. Ajabnoor MA, Mokhtar AM, Rafee AA, et al. Defective col-
lagen metabolism in Saudi patients with hernia. Ann Clin
Biochem 1992;29:430–6.
21. Cannon DJ, Read RC. Metastatic emphysema: a mechanism
for acquiring inguinal herniation. Ann Surg 1981;194:270–8.