ARTICLE OPEN ACCESS

Incidence, Prevalence, and Health Care

Outcomes in Idiopathic Intracranial
Hypertension
A Population Study
Latif Miah, MBBCh,* Huw Strafford, MSc,* Beata Fonferko-Shadrach, MPH, Joe Hollinghurst, PhD, Correspondence
Inder M.S. Sawhney, MD, Savvas Hadjikoutis, MD, Mark I. Rees, DSc, Rob Powell, PhD, Arron Lacey, PhD, and Dr. Pickrell
William O. Pickrell, PhD [email protected]

®
Neurology 2021;96:e1251-e1261. doi:10.1212/WNL.0000000000011463

Abstract
Objective
To characterize trends in incidence, prevalence, and health care outcomes in the idiopathic
intracranial hypertension (IIH) population in Wales using routinely collected health care data.

Methods
We used and validated primary and secondary care IIH diagnosis codes within the Secure
Anonymised Information Linkage databank to ascertain IIH cases and controls in a retro-
spective cohort study between 2003 and 2017. We recorded body mass index (BMI), depri-
vation quintile, CSF diversion surgery, and unscheduled hospital admissions in case and control
cohorts.

Results
We analyzed 35 million patient-years of data. There were 1,765 cases of IIH in 2017 (85%
female). The prevalence and incidence of IIH in 2017 was 76/100,000 and 7.8/100,000/y, a
significant increase from 2003 (corresponding figures = 12/100,000 and 2.3/100,000/y) (p <
0.001). IIH prevalence is associated with increasing BMI and increasing deprivation. The odds
ratio for developing IIH in the least deprived quintile compared to the most deprived quintile,
adjusted for sex and BMI, was 0.65 (95% confidence interval 0.55 to 0.76). Nine percent of IIH
cases had CSF shunts with less than 0.2% having bariatric surgery. Unscheduled hospital ad-
missions were higher in the IIH cohort compared to controls (rate ratio 5.28, p < 0.001) and in
individuals with IIH and CSF shunts compared to those without shunts (rate ratio 2.02, p < 0.01).

Conclusions
IIH incidence and prevalence is increasing considerably, corresponding to population increases
in BMI, and is associated with increased deprivation. This has important implications for health
care professionals and policy makers given the comorbidities, complications, and increased
health care utilization associated with IIH.

*These authors contributed equally to this work.

From Swansea University Medical School (L.M., H.S., B.F.-S., J.H., I.M.S.S., M.I.R., R.P., A.L., W.O.P.), Swansea University; Neurology Department (I.M.S.S., S.H., R.P., W.O.P.), Morriston
Hospital, Swansea Bay University Health Board; and Faculty of Medicine and Health (M.I.R.), University of Sydney, Australia.

Go to Neurology.org/N for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.

The Article Processing Charge was funded by the Health Data Research (HDR) UK.
This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (CC BY), which permits unrestricted use, distribution, and reproduction in any
medium, provided the original work is properly cited.

Copyright © 2021 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. e1251
 Glossary
BMI = body mass index; CI = confidence interval; GP = general practitioner; ICD-10 = International Classification of
Diseases–10; IGRP = Information Governance Review Panel; IIH = idiopathic intracranial hypertension; LSOA = lower super
output area; NSW = National Survey for Wales; SAIL = Secure Anonymised Information Linkage; WIMD = Welsh Index of
Multiple Deprivation.

Idiopathic intracranial hypertension (IIH) is a condition of definite and probable IIH cases in 2016 and 2017 from
unknown etiology that is strongly associated with obesity. IIH Morriston Hospital Swansea (the regional neurosciences
predominantly affects women of childbearing age and causes center) with SAIL data. The IIH cases were reviewed by cli-
chronic disabling headaches, visual disturbance, and, in a nicians and classed as definite if they met all IIH criteria or
minority of patients, permanent visual loss. The definitive probable if they met the criteria but a record of fundoscopy at
management is weight loss but a minority of patients require the time of diagnosis could not be found (table e-1, zenodo.
surgery in order to preserve vision.1,2 People with IIH po- org/record/4064064).8 We reviewed all patients admitted to
tentially have high rates of health care utilization due to their Morriston Hospital in 2017 with an IIH ICD-10 diagnosis
comorbidities, multiple specialist consultations, diagnostic code in their hospital admission record and checked whether
tests, CSF diversion procedures, and complications related to they had also been given a diagnosis of IIH by a consultant
CSF diversion surgery. neurologist.

Given the increasing global burden of obesity and the strong To identify patients with IIH throughout Wales, we selected
association of IIH with obesity, it would seem plausible to patients who had been given primary care or secondary care IIH
assume that the incidence of IIH, as well as associated health diagnosis codes (table e-2, zenodo.org/record/4064064) within
care utilization, is increasing.3,4 Although there is some evi- the study window (January 1, 2003, to December 31, 2017).
dence to support this,1,4 overall there is a paucity of data
regarding the epidemiology, health care utilization, and out- We excluded likely cases of secondary intracranial hypertension
comes of people with IIH. To our knowledge, there have not (CNS malignancies, venous sinus thrombosis, and malignant
been any large population-level studies of IIH in Wales. hypertension) (table e-3, zenodo.org/record/4064064). We
also excluded patients with less than 1 year of primary care data
In this retrospective cohort study, we aimed to determine the prior to diagnosis, with insufficient demographic data, and who
temporal trends of IIH incidence and prevalence in Wales and were diagnosed on their death date (figure e-1, zenodo.org/
health care utilization associated with IIH. We also investigated record/4064064). We recorded bariatric surgery procedures
the effects of socioeconomic deprivation and obesity on IIH and CSF diversion and revision procedures within the IIH co-
epidemiology. hort (see tables e-4, e-5, and e-6 for codes, zenodo.org/record/
4064064). We calculated incidence and prevalence for every
year within the study window and also calculated incidence and
Methods prevalence for each of the 7 health boards within Wales.

This was a retrospective cohort study using anonymized, We recorded deprivation using the Welsh Index of Multiple
routinely collected Welsh health care data within the Secure Deprivation (WIMD), 2011 version, in which Wales is di-
Anonymised Information Linkage databank (SAIL) at vided into 1,897 lower layer super output areas (LSOAs)
Swansea University, Wales, United Kingdom.5–7 SAIL con- containing an average of 1,600 people.9 Weighted scores from
tains anonymized linked datasets from a range of Welsh health 8 domains, representing different types of deprivation, are
sources including hospital admission and demographic data aggregated to form a WIMD score for each LSOA. The fol-
for the complete Welsh population (3.1 million, 2017 pop- lowing are the 8 domains that provide the weighted scores:
ulation estimate) and primary care records for approximately income, employment, health, education, access to services,
80% of the Welsh population. SAIL uses an established and housing, community safety, and physical environment. Each
validated split-file approach for anonymized data linkage.5,6 LSOA in Wales has been ranked from most deprived to least
deprived according to its WIMD score and then grouped into
In the United Kingdom, primary care consists primarily of quintiles, with quintile 1 being the most deprived and quintile
general practitioners (GPs). Patients with IIH can first pre- 5 being the least deprived.
sent to their GP, optometry, neurology, or emergency de-
partment before being given a diagnosis by ophthalmologists We created a control cohort for the case (IIH) cohort by 3:1
or neurologists in secondary care. matching on sex, age (week of birth), and WIMD quintile at
time of diagnosis of IIH. For any analysis of the control cohort
To validate the accuracy of IIH diagnosis codes in routinely involving diagnosis date, the control was given the diagnosis
collected data, we anonymously linked a list of successive date of the matched patient with IIH.

e1252 Neurology | Volume 96, Number 8 | February 23, 2021 Neurology.org/N

To measure obesity within the IIH and control cohorts, we anonymized, routinely collected data and therefore written in-
used primary care body mass index (BMI) data measured formed consent was not required. The Research Ethics Service
during routine GP consultations. BMI is defined as body mass has previously confirmed that SAIL projects using anonymized,
in kilograms divided by height in meters squared. Within the routinely collected data do not require specific UK National
IIH cohort, most had at least 1 recorded BMI. For those with Health Service research ethics committee approval.
more than 1 recorded BMI, we used the BMI nearest to the
diagnosis date. We excluded likely erroneous BMI values of Data Availability
less than 10 kgm−2 and greater than 70 kgm−2. The detailed anonymized patient data used in this study are
potentially re-identifiable and therefore not directly available
We used BMI data recorded in primary care (GP) and self- for sharing but are available within the SAIL Databank at
reported BMI data from the National Survey for Wales Swansea University. All proposals to use SAIL data are subject
(NSW) to measure temporal trends in obesity. The NSW is to review by an IGRP. Before any data can be accessed, ap-
completed on behalf of the Welsh Government and over proval must be given by the IGRP. The IGRP gives careful
11,000 randomly selected individuals annually report on a consideration to each project to ensure proper and appro-
range of health- and lifestyle-related issues.10 Both sources of priate use of SAIL data. When access has been approved, it is
BMI data are open to ascertainment bias. People with low or gained through a privacy-protecting safe haven and remote
high BMI are more likely to have their BMI recorded by their access system referred to as the SAIL Gateway. SAIL has
GP and questionnaire respondents are likely to underestimate established an application process to be followed by anyone
their weight. We therefore also used a mean of both GP and who would like to access data via SAIL (saildatabank.com/
NSW BMI data to estimate temporal trends in obesity (BMI application-process).
>30 kg/m2) rates in the Welsh population.

We recorded the rate of unscheduled hospital admissions in Results

the IIH and control cohort. Unscheduled hospital admissions We analyzed a total of 35 million patient-years of data (a mean
were defined as emergency department attendances or of 2.3 million patient-years of data per year) during the study
emergency/unscheduled hospital admissions (Patient Epi- period (2003–2017).
sode Dataset for Wales dataset).
Validating Diagnosis Codes
Statistical Methods We identified 153 successive individuals with an IIH diagnosis
We calculated IIH prevalence by dividing the number of admitted to our neuroscience center in 2016 and 2017 (see
people with a diagnosis of IIH by the total number of people, Methods). In order to validate the accuracy of IIH diagnosis
with sufficient demographic data, registered with a GP within codes used in the routinely collected data, we anonymously
SAIL on July 1 each year. We calculated annual IIH incidence linked these 153 individuals with IIH diagnoses (138 definite
rates by dividing the number of new IIH diagnosis by the and 15 probable) to health care data within SAIL. A total of
number of patient-years at risk in that time period. 149 of these cases had linked primary and secondary care data;
all 153 had linked secondary care data. We were able to as-
We used a logistic regression model to compare the relative certain 125 out of our 149 individuals with definite or prob-
contributions of BMI and deprivation on the risk of de- able IIH using GP codes alone (sensitivity 84%), 130/153
veloping a new diagnosis of IIH. In this model, we included (sensitivity 85%) using hospital codes alone, and 140/153
every person with a primary care BMI measurement within using a combination of GP and hospital codes (sensitivity
the study period. A diagnosis of IIH within the study period 92%). A total of 141 out of all 153 individuals having a hospital
was the binary outcome variable. For controls with multiple ICD-10 code for IIH admitted to a neuroscience center in
BMI measurements, we used the last recorded BMI before the 2017 had a consultant neurologist diagnosis of IIH (specificity
end of the study period. For individuals with IIH, we used the 87%; table e-7, zenodo.org/record/4064064). We used a
BMI and deprivation quintile closest to diagnosis. We ex- combination of GP and hospital codes to ascertain cases with
cluded any person who died during the study window. IIH for the remainder of the study.

We used a paired t test to compare changes in BMI mea- IIH Prevalence, IIH Incidence, and
surements for individuals, proportion tests to compare pro- Obesity Trends
portions, and rate ratio tests to compare rates of hospital There were 2,275 people with primary or secondary care IIH
admission. We used R version 3.5.3 for statistical analysis. diagnosis codes within the study window. After excluding
potential secondary causes of intracranial hypertension and
Standard Protocol Approvals, Registrations, cases with no demographic data or no preceding primary care
and Patient Consents data, there was a prevalent IIH cohort of 1,765 in 2017 (figure
All studies using SAIL data need independent Information e-1, zenodo.org/record/4064064; and table 1). A total of 86%
Governance Review Panel (IGRP) approval. This study of the cohort had a GP BMI recording at or after diagnosis of
obtained IGRP approval (ref 0695). This study used IIH (figure 1).

Neurology.org/N Neurology | Volume 96, Number 8 | February 23, 2021 e1253

 Table 1 Characteristics of the Study Cohort, n (%) Table 1 Characteristics of the Study Cohort, n (%) (continued)
All Male Female CSF shunts Control cohort

No. of cases 1,765 (100) 272 (15.4) 1,493 (84.6) Mean (SD) 35.7 (7.3) 26.6 (6.5)

Age at diagnosis, y WIMD deprivation quintile

<18 270 (15.3) 88 (32.6) 182 (12.2) 1 (most deprived) 54 (34.2) 1,560 (29.5)

18–30 743 (42.1) 49 (18.0) 694 (46.5) 2 42 (26.6) 1,227 (23.2)

30–40 390 (22.1) 35 (12.9) 355 (23.8) 3 32 (20.3) 1,023 (19.3)

40–50 188 (10.7) 31 (11.4) 157 (10.5) 4 16 (10.1) 786 (14.8)

50–60 93 (5.3) 28 (10.3) 65 (4.4) 5 (least deprived) 14 (8.9) 699 (13.2)

60+ 81 (4.6) 41 (15.1) 40 (2.7) Mean (SD) 2.33 (1.28) 2.59 (1.39)

Age, y, mean (SD) 30.1 (14.2) 33.6 (21.3) 29.4 (12.4) Acetazolamide prescribed 116 (73.4) 6 (0.1)
2
BMI, kg/m
Abbreviations: BMI = body mass index; WIMD = Welsh Index of Multiple
Deprivation.
Number with 1,522 (86) 183 (67.3) 1,339 (89.7) a
Because of small numbers of patients with CSF shunts in the 50–60 and 60+
BMI recorded age groups, we have reported these together as patients with CSF shunts
>50 years old.
<25 164 (10.8) 33 (18.0) 131 (9.8)

25–30 284 (18.7) 66 (36.1) 218 (16.3)

>30 1,074 (70.6) 84 (46.0) 990 (74.0) The prevalence of IIH increased greatly from 12/100,000 in
2003 to 76/100,000 in 2017 (p < 0.001) (figure 2A). The
Mean (SD) 35.2 (8.5) 31.5 (8.4) 35.7(8.4)
incidence of IIH increased greatly from 2.3/100,000/y in
WIMD deprivation quintile 2003 to 7.8/100,000/y in 2017 (p < 0.001) (figure 2B). The
1 (most deprived) 520 (29.5) 68 (25.0) 452 (30.3)
proportion of obese individuals (BMI >30 kg/m2) in Wales,
as measured using primary care data, increased significantly,
2 409 (23.2) 66 (24.3) 343 (23.0)
with 29% of the population being obese in 2003 compared to
3 341 (19.3) 62 (22.8) 279 (18.7) 40% in 2017 (p < 0.001). There were also significant increases
in obesity as measured by questionnaire data (figure 2C). IIH
4 262 (14.8) 40 (14.7) 222 (14.9)
incidence and prevalence by Welsh health boards is shown in
5 (least deprived) 233 (13.2) 36 (13.2) 197 (13.2) figure 3 and table e-8 (zenodo.org/record/4064064).
Mean (SD) 2.59 (1.39) 2.67 (1.34) 2.58 (1.39)
Association Among IIH, Deprivation, and BMI
Acetazolamide prescribed 965 (54.7) 89 (32.7) 876 (58.7) Figure 4 illustrates the association between both IIH preva-
CSF shunts Control cohort lence and incidence with obesity and the association between
obesity and deprivation. Figure 4, A–D, use incidence and
Number of cases 158 (9.0) 5,295 (100)
prevalence averaged over the whole study period and figure
Age at diagnosis, y 4E uses obesity data for 2015 only. Similar patterns are seen
<18 23 (14.6) 810 (15.3) for all years in the study period (figures e-2, e-3, and e-4;
zenodo.org/record/4064064).
18–30 70 (44.3) 2,229 (42.1)

30–40 43 (27.2) 1,170 (22.1) For both men and women, IIH prevalence and incidence are
40–50 16 (10.1) 564 (10.7)
strongly associated with BMI. For obese women (BMI >30 kg/m2),
the mean prevalence and incidence of IIH was 180/100,000 and
50–60 6 (3.8) (>50)a 279 (5.3) 23.5/100,000/y. The corresponding figures for women with an
60+ 243 (4.6) ideal BMI (20 kg/m2 < BMI < 25 kg/m2) were 13.2/100,000 and
1.6/100,000/y, respectively. For obese men, the mean prevalence
Age, y, mean (SD) 27.6(11.6) 30.1 (14.2)
and incidence of IIH was 21.2/100,000 and 2.6/100,000/y. The
BMI, kg/m2 corresponding figures for men with an ideal BMI were 7.6/100,000
Number with BMI recorded 140 (88.6) 4,251 (80.3) and 1.6/100,000/y, respectively. In 2015, 35.9% of people in the
most deprived quintile were obese, compared to 24.4% in the least
<25 10 (7.1) 2015 (47.4)
deprived quintile.
25–30 21 (15.0) 1,196 (28.1)

>30 109 (77.9) 1,040 (24.5)

Table 2 shows odds ratios for developing IIH based on sex,
BMI, and deprivation score.

e1254 Neurology | Volume 96, Number 8 | February 23, 2021 Neurology.org/N

 Patients with IIH who had CSF diversion surgery (n = 158)
Figure 1 Body Mass Index (BMI) at Diagnosis for the Idio- had 1,215 unscheduled hospital visits over 407,751 days
pathic Intracranial Hypertension (IIH) Cohort (mean rate of 1.088 visits/patient/year). When compared to
patients with IIH without CSF shunts, the rate ratio for un-
scheduled hospital visits was 2.02 (95% CI 1.89 to 2.15; p <
0.01) (figure 5).

Discussion
We used a combination of primary and secondary care di-
agnosis codes in anonymized, routinely collected data to as-
certain cases of IIH in this large, retrospective cohort study.
Uniquely, we were able to validate our IIH case ascertainment
method and found a sensitivity of 92% when identifying a
linked set of confirmed IIH cases. Diagnostic codes used in
the United Kingdom are considered to be accurate,11 but
specific validation of UK primary and secondary IIH diagnosis
codes, which we have performed here, is rare. It would seem
reasonable to use similar methods in future epidemiologic
studies.
The distribution of BMI measurements at nearest IIH diagnosis for women
(ochre) and men (blue). The y-axis shows the proportion of either all female
or all male cases. We found a significant increase in IIH incidence and preva-
lence in Wales. The prevalence of IIH in Wales increased
sixfold from 12/100,000 in 2003 to 76/100,000 in 2017 and
Outcome Data the incidence of IIH increased threefold from 2.3/100,000/y
Thirteen patients with IIH (0.78%) were recorded as being in 2003 to 7.8/100,000/y in 2017. There were corresponding
blind a mean of 858 days (28 months) after diagnosis. A total increases in obesity rates, using primary care BMI measure-
of 32 patients (1.9%) were recorded as having at least mod- ments, with 29% of the population being obese in 2003
erate visual impairment a mean of 804 days (26 months) after compared to 40% in 2017.
diagnosis. A total of 158 patients (9%) underwent CSF di-
version procedures a mean of 491 days (16 months) after Until recently, there has been a paucity of large-scale epide-
diagnosis (see table 1 for the characteristics of the shunt co- miologic studies of IIH. Older studies have found the in-
hort and figure e-5 [zenodo.org/record/4064064] for a cidence of IIH to be 1–2/100,000.12–14 A recent meta-analysis
Kaplan-Meier plot). A total of 70 (44%) patients who had of 15 studies, from 10 different countries, included 889 patients
CSF diversion procedures went on to have at least 1 revision with IIH and found a pooled incidence of 1.2/100,000/y.15
surgery. Larger, more recent studies have found higher, and increasing,
incidence rates of IIH, in line with our findings. A study using
A total of 691 patients with IIH had at least 2 BMI readings routinely collected English secondary care data identified
after their diagnosis (first BMI was within 1 year of diagnosis 23,182 new cases of IIH (82.4% female) with an incidence of
and the latest BMI reading was a mean of 5.1 years later). 2.3/100,000/y in 2002 and 4.7/100,000/y in 2016.1 Another
These patients’ BMI increased by 0.48 kg/m2 (p = 0.02, 95% recent study used English primary care data to study cardio-
confidence interval [CI] 0.07 to 0.89). Fewer than 6 optic vascular risk in women with IIH and found that the incidence of
nerve fenestration procedures or gastric banding procedures IIH in women increased from 2.5 in 2005 to 9.3/100,000/y in
were performed (note that due to SAIL anonymization 2017.16 A Minnesota study also demonstrated over a 24-year
guidance we cannot report groups of fewer than 6 individuals period that IIH incidence was increasing in a highly correlated
or events). fashion with obesity.4

Health Care Utilization The considerable increase in IIH incidence is multifactorial

During the study period, the IIH cohort had 4,818 un- but likely predominately due to rising obesity rates. The
scheduled hospital visits over 3,262,942 days (mean rate of worldwide prevalence of obesity nearly tripled between 1975
0.54/patient/year), considerably more than the 2,755 un- and 2016 and therefore these results also have global
scheduled hospital visits over 9,860,456 days in the control relevance.17,18 The increase in IIH incidence may also be
cohort (mean rate of 0.10/patient/year; rate ratio 5.28; 95% attributable to increased IIH diagnosis rates due to raised
CI 5.04 to 5.54; p < 0.001). This amounts to 777 additional awareness of the condition and greater use of digital fundu-
unscheduled hospital visits for the cohort of 1,765 patients scopy at routine optometry appointments. Misdiagnosis of
with IIH when compared to controls (figure 5). IIH may also contribute to increasing IIH incidence; for

Neurology.org/N Neurology | Volume 96, Number 8 | February 23, 2021 e1255

 Figure 2 Prevalence and Incidence of Idiopathic Intracranial Hypertension (IIH) Compared With Obesity (2003–2017)

Changes in (A) IIH prevalence and (B) IIH incidence with time during the study period. Ochre lines represent the female population, blue lines represent the
male population, and green lines represent the combined population. (C) Changes in the proportion of the population who are obese (body mass index >30
kg/m2) from general practitioner (GP) BMI measurements (red line), National Survey for Wales (NSW) data (green line), and a mean average of the two (blue
line) during the study period. The x-axis in all graphs represents time in years. Note NSW data were not available for 2003–2005 or 2015–2017.

example, there can be issues when interpreting fundoscopy in The association between IIH and deprivation can be
patients with headache.19 explained by increasing obesity rates in more deprived areas.
However, results from our logistic regression model suggest
We found a strong association between increasing BMI, sex that even when adjusting for BMI, there was still an associa-
(female), and IIH. Around 85% of our IIH cohort was female, tion between IIH and deprivation in women, but not men.
similar to other studies.1,19 We also found a significant association This result was also seen in a recent study of English women
with increased deprivation and IIH, particularly in women. A total with IIH.5 We found that women were 1.5 times more likely
of 52.6% of our IIH cohort came from the 2 most deprived to develop IIH in the most deprived compared to the least
quintiles, a very similar finding to a recent English study that had deprived areas, even after adjusting for BMI. In women it
52.3% of its cases from the 2 most deprived quintiles.1 therefore seems likely that factors that are associated with

e1256 Neurology | Volume 96, Number 8 | February 23, 2021 Neurology.org/N

Figure 3 Choropleths Comparing Welsh Idiopathic Intracranial Hypertension (IIH) Incidence and Prevalence With
Deprivation

(A) Map of the current Welsh health boards. Choropleth maps show the distribution of (B) deprivation, measured as a percentage of the health board
population in the 2 most deprived Welsh Index of Multiple Deprivation (WIMD) quintiles (1–2); (C) IIH incidence 2015–2017 by the health board; and (D) IIH
prevalence in 2017 by the health board. Data for the Powys Health board could not be presented for incidence and prevalence due to restrictions in obtaining
small numbers from Secure Anonymised Information Linkage. See also table e-8 (zenodo.org/record/4064064). NHS = National Health Service.

deprivation, apart from obesity alone, also contribute to the the evidence that IIH in men has different characteristics than
etiology of IIH. Central adiposity is prominent in IIH and it IIH in women.1
may be that the distribution of fat changes with deprivation.20
Factors associated with deprivation such as diet, pollution, We have demonstrated that people with IIH have increased
smoking, and psychosocial stress may have a metabolic or en- rates of unscheduled health care utilization compared with a
docrine effect by, for example, increasing circulating androgen matched control cohort. The rate ratio for unscheduled
levels, which are associated with IIH and altered CSF flow.21 hospital admissions in the IIH cohort compared to the con-
trols was 5.28 (95% CI 5.04 to 5.54). A considerable pro-
We found that for men, although our numbers were smaller, portion of this excess in unscheduled hospital admissions
IIH is associated with BMI only and not deprivation. There occurs at the time of diagnosis and can be explained by the
was a weaker association of BMI with IIH and a smaller in- need for urgent investigation of papilledema with brain im-
crease in IIH incidence with time in men when compared with aging and spinal fluid analysis. However, there is also a con-
women. Men are 4 times less likely to develop IIH than siderable excess in unscheduled hospital admissions up to 2
women after adjusting for BMI and deprivation. This adds to years after diagnosis.

Neurology.org/N Neurology | Volume 96, Number 8 | February 23, 2021 e1257

 Figure 4 Relationship of Idiopathic Intracranial Hypertension (IIH) Prevalence and Incidence With Obesity and Deprivation

Association between IIH prevalence (A) and incidence (B) with body mass index (BMI). BMI <20 kg/m2 is defined as underweight, 20 kg/m2 < BMI < 25 kg/m2
normal weight, 25 kg/m2 < BMI < 30 kg/m2 overweight, and BMI >30 kg/m2 as obese. (C, D) Relationship between IIH prevalence and incidence with deprivation
as measured by the Welsh Index of Multiple Deprivation (1 = most deprived, 5 = least deprived). (E) Relationship between obesity and deprivation for 2015
(similar relationships exist for other study years; see figure e-2, zenodo.org/record/4064064). The green line is obtained from primary care BMI measure-
ments, the blue line from Welsh Health Survey BMI data, and the red line is the mean of the 2 measurements. GP = general practitioner; NSW = National
Survey for Wales.

It is likely that these admissions are for severe headache. In the cases.22 There is therefore some scope to reduce emergency
past, emergency admissions may have led to therapeutic admissions through better management of headache, patient
lumbar punctures, although recent UK consensus guidelines education, and rapid access to outpatient specialist advice.
advise against therapeutic lumbar punctures in the majority of Interestingly, the rate of unscheduled admissions is higher in

e1258 Neurology | Volume 96, Number 8 | February 23, 2021 Neurology.org/N

Table 2 Odds Ratios (ORs) (95% Confidence Intervals Figure 5 Rates of Unscheduled Hospital Visits Among 3
[CIs]) for Developing Idiopathic Intracranial Cohorts
Hypertension by Sex, Deprivation, and Body
Mass Index (BMI) Based on a Logistic Regression
Model
Model OR (95% CI) p Value

Sex

Male 1 (reference)

Female 3.96 (3.51–4.50) <0.00001a

Deprivation quintile

1 (most deprived) 1 (reference)

2 0.88 (0.76–1.01) 0.068

3 0.79 (0.68–0.91)a 0.0015a

4 0.73 (0.62–0.86)a 0.0001a

5 (least deprived) 0.65 (0.55–0.76)a <0.00001a

Rates of unscheduled hospital visits (per person per year) versus time since
BMI 1.13 (1.13–1.14)a <0.00001a
idiopathic intracranial hypertension (IIH) diagnosis in months. A negative
value indicates time before the diagnosis and a positive value indicates time
Male after the diagnosis. Green line represents the IIH cohort; red line represents
the control cohort (matched on age, sex, and deprivation quintile); blue line
Deprivation quintile represents the IIH cohort with CSF diversion surgery (CSF shunts).

1 (most deprived) 1 (reference)

2 1.29 (0.86–1.94) 0.23

have significantly increased unscheduled health care admis-
3 0.97 (0.62–1.51) 0.90
sion rates compared with individuals with IIH who have not
4 0.74 (0.45–1.21) 0.24 undergone CSF diversion procedures (rate ratio for un-
5 (least deprived) 0.76 (0.46–1.22) 0.26
scheduled admissions 2.02, 95% CI 1.89 to 2.15). They have
around a 40% chance of having at least 1 CSF shunt revision
BMI 1.10 (1.08–1.12)a <0.00001a
procedure.
Female
Forty percent of our IIH cohort had a second BMI recording
Deprivation quintile
after diagnosis. In these 691 individuals, their mean BMI in-
1 (most deprived) 1 (reference) creased by 0.48 kgm−2 (p = 0.02, 95% CI 0.07 to 0.89). This
2 0.84 (0.72–0.97)a 0.019a corresponds to an increase of 1.3 kg (using an average height
of 1.62 meters for women in Wales). This reflects the diffi-
3 0.77 (0.65–0.90)a 0.001a
culty in body weight reduction despite weight loss being the
4 0.73 (0.61–0.86)a 0.0003a main management strategy for IIH.2 Despite this, there were
5 (least deprived) 0.65 (0.54–0.77)a <0.00001a
very low rates of bariatric surgery in our cohort. Although 9%
had shunts, less than 0.3% had bariatric surgery to tackle the
BMI 1.13 (1.13–1.14)a <0.00001a
underlying obesity. Moreover, the shunt population still had
Deprivation is measured using the Welsh Index of Multiple Deprivation
high rates of hospital attendance despite invasive treatment.
(WIMD) with 1 being the most deprived quintile and 5 being the least de-
prived quintile (see methods for more details). Results are displayed for 3
models: 1) male and female; 2) male only 3) female only. All individuals in We analyzed a large number (32 million patient-years) of
Wales with available BMI measurements during the study period were in- population-level data, integrating both primary and secondary
cluded (1,522 with IIH and 1,655,014 without).
a
p Value <0.05. care datasets and obtaining a large cohort of IIH cases and
matched controls. Using routinely collected data reduces re-
cruitment bias seen in other studies. We also include routinely
the IIH cohort in the 3 years leading up to diagnosis, sug- collected BMI measurements for the majority of our IIH cohort
gesting an opportunity for earlier diagnosis and earlier in addition to deprivation measures in our regression model.
intervention.
This was a retrospective study. Routinely collected health care
About 9% of people with IIH in Wales receive CSF diversion data are not primarily collected for research purposes, can be
procedures a mean of 1.33 years after diagnosis. Individuals incomplete, and may contain inaccurate diagnosis codes. At
with IIH who have undergone CSF diversion procedures also the time of analysis, within SAIL we had access to 100% of the

Neurology.org/N Neurology | Volume 96, Number 8 | February 23, 2021 e1259

 Welsh population’s secondary care diagnosis codes but only
80% of the population’s primary care data. BMI is not rou- Appendix Authors
tinely measured by GPs, raising the possibility of ascertain- Name Location Contribution
ment bias as those who have BMI recorded are more likely to
Latif Miah, Swansea University Drafted and edited the
have ongoing weight issues. We made an attempt to adjust for MBBCh Medical School manuscript, collected and
this by also looking at national questionnaire data for BMI, analysed the data
but these self-reported data are subject to underreporting of
Huw Swansea University Analysed the data,
obesity. We used an area-based measure of deprivation Strafford, Medical School performed statistical
(WIMD), which, like other commonly used deprivation MSc analysis and edited the
manuscript
markers, does not take into account individual levels of dep-
rivation. We did not specifically record prescriptions of drugs Beata Swansea University Collected data, edited the
Fonferko- Medical School manuscript
such as tetracyclines and retinoids that can cause secondary Shadrach,
intracranial hypertension although this probably accounts for MPH
a small number of cases.23 Joe Swansea University Provided statistical advice,
Hollinghurst, Medical School edited the manuscript
PhD
IIH incidence and prevalence in Wales are increasing con-
siderably, corresponding to population increases in BMI. IIH Inder M.S. Neurology Department, Provided supervision,
Sawhney, MD Morriston Hospital, edited the manuscript
is associated with increasing deprivation in women even after Swansea Bay University
adjusting for obesity, suggesting additional etiologic factors Health Board
associated with deprivation apart from BMI. This effect was
Savvas Neurology Department, Provided specialist IIH
not seen in men, pointing to sex-specific drivers for IIH. Hadjikoutis, Morriston Hospital, advice, edited the
MD Swansea Bay University manuscript
Health Board
The increasing incidence of IIH, together with the increased
health care utilization in individuals with IIH and particularly Mark I. Rees, Faculty of Medicine and Provided project
DSc Health, University of investment, provided
those who have CSF shunts, have important implications for Sydney, Australia supervision, edited the
health care professionals and policy makers. manuscript

Rob Powell, Neurology Department, Provided supervision,

Acknowledgment PhD Morriston Hospital, edited the manuscript
This study makes use of anonymized data held in the Secure Swansea Bay University
Health Board
Anonymised Information Linkage (SAIL) system, which is
part of the national e-health records research infrastructure for Arron Lacey, Swansea University Provided supervision,
PhD Medical School assisted with data
Wales. The authors thank the data providers who make analysis, edited the
anonymized data available for research. manuscript

William Neurology Department, Designed, coordinated

Owen Morriston Hospital, and supervised the study,
Study Funding Pickrell, PhD Swansea Bay University drafted and edited the
The study was not directly funded but was supported by the Health Board manuscript
Brain Repair and Intracranial Neurotherapeutics (BRAIN)
Unit and the Wales Gene Park, which are funded by Health
and Care Research Wales. This work was also supported by
Health Data Research UK, which receives its funding from References
1. Mollan SP, Aguiar M, Evison F, Frew E, Sinclair AJ. The expanding burden of
HDR UK Ltd. (HDR-9006) funded by the UK Medical Re- idiopathic intracranial hypertension. Eye 2019;33:478–485.
search Council, Engineering and Physical Sciences Research 2. Sinclair AJ, Burdon MA, Nightingale PG, et al. Low energy diet and intracranial
pressure in women with idiopathic intracranial hypertension: prospective cohort
Council, Economic and Social Research Council, Department study. BMJ 2010;341:c2701.
of Health and Social Care (England), Chief Scientist Office of 3. Wang YC, McPherson K, Marsh T, Gortmaker SL, Brown M. Health and economic
burden of the projected obesity trends in the USA and the UK. Lancet 2011;378:
the Scottish Government Health and Social Care Director- 815–825.
ates, Health and Social Care Research and Development 4. Kilgore KP, Lee MS, Leavitt JA, et al. Re-evaluating the incidence of idiopathic intracranial
hypertension in an era of increasing obesity. Ophthalmology 2017;124:697.
Division (Welsh Government), Public Health Agency 5. Ford DV, Jones KH, Verplancke JP, et al. The SAIL Databank: building a national
(Northern Ireland), British Heart Foundation (BHF), and the architecture for e-health research and evaluation. BMC Health Serv Res 2009;4:157.
Wellcome Trust. 6. Lyons RA, Jones KH, John G, et al. The SAIL databank: linking multiple health and
social care datasets. BMC Med Inform Decis Mak 2009;9:3.
7. SAIL Databank: the Secure Anonymised Information Linkage Databank [Online].
Disclosure Available at: saildatabank.com/. Accessed August 15, 2019.
8. Friedman DI, Liu GT, Digre KB. Revised diagnostic criteria for the pseudotumor
The authors report no relevant disclosures. Go to Neurology. cerebri syndrome in adults and children. Neurology 2013;81:1159–1165.
org/N for full disclosures. 9. Welsh Government. Welsh Index of Multiple Deprivation. 2020. Available at: https://
gov.wales/welsh-index-multiple-deprivation. Accessed January 12, 2021.
10. Welsh Government. Welsh Health Survey. 2020. Available at: statswales.gov.wales/
Publication History Catalogue/Health-and-Social-Care/Welsh-Health-Survey. Accessed February 16,
2020.
Received by Neurology July 22, 2020. Accepted in final form 11. Khan NF, Harrison SE, Rose PW. Validity of diagnostic coding within the general
October 23, 2020. practice research database: a systematic review. Br J Gen Pract 2010;60:e128–e136.

e1260 Neurology | Volume 96, Number 8 | February 23, 2021 Neurology.org/N

12. Durcan FJ, Corbett JJ, Wall M. The incidence of pseudotumor cerebri: population 18. Public Health Wales Observatory. Obesity in Wales [Internet]. 2019. Available at:
studies in Iowa and Louisiana. Arch Neurol 1998;45:875–877. publichealthwalesobservatory.wales.nhs.uk/obesityinwales. Accessed December
13. Radhakrishnan K, Ahlskog JE, Cross SA, Kurland LT, O’Fallon WM. Idiopathic 2019.
intracranial hypertension (pseudotumor cerebri): descriptive epidemiology in 19. Fisayo A, Beau BB, Newman NJ, Biousse V. Overdiagnosis of idiopathic intracranial
Rochester, Minn, 1976 to 1990. Arch Neurol 1993;50:78–80. hypertension. Neurology 2016;86:341–350.
14. Raoof N, Sharrack B, Pepper IM, Hickman SJ. The incidence and prevalence of idiopathic 20. Hornby C, Botfield H, O’Reilly MW, et al. Evaluating the fat distribution in idiopathic
intracranial hypertension in Sheffield, UK. Eur J Neurol 2011;18: 1266–1268. intracranial hypertension using dual-energy X-ray absorptiometry scanning. Neuro-
15. McCluskey G, Doherty-Allan R, McCarron P, et al. Meta-analysis and systematic ophthalmology 2017;42:99–104.
review of population-based epidemiological studies in idiopathic intracranial hyper- 21. O’Reilly MW, Westgate CS, Hornby C, et al. A unique androgen excess signature in
tension. Eur J Neurol 2018;25:1218–1227. idiopathic intracranial hypertension is linked to cerebrospinal fluid dynamics. JCI
16. Adderley N, Subramanian A, Nirantharakumar K, et al. Association between idio- Insight 2019;4:e125348.
pathic intracranial hypertension and risk of cardiovascular diseases in women in the 22. Mollan SP, Davies B, Silver NC, et al. Idiopathic intracranial hypertension: consensus
United Kingdom. JAMA Neurol 2019;76:1088. guidelines on management. J Neurol Neurosurg Psychiatry 2018;89:1088–1100.
17. World Health Organization. Obesity and overweight [Internet]. 2019. Available at: who.int/ 23. Crum OM, Kilgore KP, Sharma R, et al. Etiology of papilloedema in patients in the eye
news-room/fact-sheets/detail/obesity-and-overweight. Accessed December 2019. clinic setting. JAMA Netw Open 2020;3:e206625.

Neurology.org/N Neurology | Volume 96, Number 8 | February 23, 2021 e1261

 Incidence, Prevalence, and Health Care Outcomes in Idiopathic Intracranial
Hypertension: A Population Study
Latif Miah, Huw Strafford, Beata Fonferko-Shadrach, et al.
Neurology 2021;96;e1251-e1261 Published Online before print January 20, 2021
DOI 10.1212/WNL.0000000000011463

This information is current as of January 20, 2021

Updated Information & including high resolution figures, can be found at:
Services http://n.neurology.org/content/96/8/e1251.full

References This article cites 18 articles, 4 of which you can access for free at:
http://n.neurology.org/content/96/8/e1251.full#ref-list-1
Citations This article has been cited by 1 HighWire-hosted articles:
http://n.neurology.org/content/96/8/e1251.full##otherarticles
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Cohort studies
http://n.neurology.org/cgi/collection/cohort_studies
Idiopathic intracranial hypertension
http://n.neurology.org/cgi/collection/idiopathic_intracranial_hypertensi
on
Incidence studies
http://n.neurology.org/cgi/collection/incidence_studies
Prevalence studies
http://n.neurology.org/cgi/collection/prevalence_studies
Quality of life
http://n.neurology.org/cgi/collection/quality_of_life
Permissions & Licensing Information about reproducing this article in parts (figures,tables) or in
its entirety can be found online at:
http://www.neurology.org/about/about_the_journal#permissions
Reprints Information about ordering reprints can be found online:
http://n.neurology.org/subscribers/advertise

Neurology ® is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright Copyright © 2021 The Author(s). Published by
Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology.. All rights reserved. Print
ISSN: 0028-3878. Online ISSN: 1526-632X.