Congenital and infantile hydrocephalus

The epidemiology of congenital and infantile hydrocephalus has been explored in several studies. In a Swedish epidemiological study, the mean crude prevalence was 0.60 per 1000 live births in the period 1973–1990 [1], excluding cases due to neural tube defects. This compares well to a population-based retrospective American cohort from 1991 to 2000, where the prevalence of congenital hydrocephalus was 0.59 per 1000 [2]. Data from four European registries of congenital malformations also approximate this, showing a prevalence of 0.47 per 1000 children born in the period 1996–2003 [3]. A Danish study including data from a local register of congenital malformations from the period 1986–1998 showed a prevalence of 0.4 per 1000 births, but 8 out of the 29 cases were stillbirths [4].

As to time trends of congenital hydrocephalus, the Swedish study showed a significant increase in prevalence from 1973–78 to 1979–82, considered due to the enhanced survival of very and extremely preterm infants [1]. However, this increase did not continue from 1983 to 1990, indicating an improved outcome in preterm infants in this period. In line with this trend, an Italian single-institution study showed a significant 8.8% decrease of the incidence between the periods 1985–90 and 2000–05 [5], and an epidemiological study from China showed a significant downward trend in the incidence of congenital hydrocephalus from 1998 to 2009 [6].

Idiopathic normal pressure hydrocephalus (iNPH)

The epidemiology of iNPH varies widely between the published studies, probably reflecting both inconsistent definitions of iNPH, differences between the populations sampled, and differences in the methodological approach towards the populations in question. As we shall see, many studies on the epidemiology of NPH do not differentiate between idiopathic and secondary NPH, something that makes the results even more difficult to evaluate. An overview of some studies is given in Table 5.1.

Table 5.1 Occurrence of NPH in some epidemiological studies

In the earliest epidemiological study, Casmiro et al. examined five age cohorts over the age of 65 in the state of San Marino in 1985 [7]; 396 inhabitants out of a sample of 488 (81%) participated. The authors based the diagnosis on (1) absence of known causes of secondary NPH, (2) impaired gait, and (3) CT scans showing findings indicative of NPH. They found two cases of NPH, yielding a prevalence of 0.5%.

Vanneste et al. conducted a retrospective study in 1992, including only patients admitted to neurosurgical centers and having a good clinical response from treatment with shunt insertion [8]. This yielded an incidence of 2.2 per million inhabitants per year. However, the authors did not differentiate between idiopathic and secondary NPH, and the prevalence could also not be estimated from this design.

In a study from 1995 by Trenkwalder et al., 982 out of a total population of 1192 inhabitants over the age of 65 years in two German villages were examined primarily for parkinsonism [9]; 45 had clinical signs of parkinsonism. Out of these, 4 patients were diagnosed as having NPH, making the prevalence of NPH 0.41%. The diagnostic criteria in this study included “typical gait disorder, bladder incontinence, dementia to a variable extent, and corresponding periventricular lucencies with ventricular enlargement.” The authors did not distinguish between idiopathic and secondary NPH.

Vale and Miranda reported in 2002 the features of 168 patients with dementia, all admitted to a tertiary-care outpatient clinic in Brazil from 1997 to 1999 [10]. The patients were examined 18 months after their initial assessment. All patients underwent CT and/or MRI, but no results from these examinations relevant for the diagnosis of NPH are stated. The authors reported that 10 patients (5.4%) had NPH. Neither diagnostic nor clinical criteria for NPH are defined, and the authors do not distinguish between idiopathic and secondary NPH.

Krauss and Halve in 1997 conducted a questionnaire survey of 82 neurosurgical centers in Germany, out of which 65% responded [11]. The estimated annual incidence of NPH was 1.8/100 000. It is not clearly defined in the questionnaire whether shunt insertion and shunt response was necessary to confirm the diagnosis, and the diagnostic criteria varied from center to center. The authors estimated that “about two thirds of patients suffered from iNPH, and one third from sNPH,” but no further data are given on this point.

As part of a local epidemiology project, Knopman et al. ascertained retrospectively incident cases of dementia in Rochester, Minnesota from 1990 to 1994 [12]. They found 560 cases of dementia. In 5 out of these, NPH was initially suspected. 3 underwent shunt placement, but none responded to therapy.

In contrast to this study stands a retrospective study among residents of assisted-living and extended-care facilities in four nursing homes [13]. Using a retrospective chart analysis, the authors included 147 patient records and found an estimated incidence of suspected iNPH ranging from 9% to 14%, depending on the diagnostic criteria used.

The authors of this chapter prospectively explored the epidemiology of iNPH in a Norwegian county of 220 000 inhabitants, by actively informing the public and professional health workers about the condition, asking for referral of suspected individuals on a broad clinical basis [14]. This yielded 86 referrals, of which 63 fulfilled the diagnostic criteria for “probable iNPH” [15] and 48 fulfilled the diagnostic criteria for “possible iNPH” [15]. From this, the prevalence of “probable iNPH” could be calculated as 21.9/100 000, with an incidence of 5.5/100 000/year. The prevalence of “possible iNPH” could be calculated as 28.7/100 000, with an incidence of 7.3/100000/year. Most importantly, the prevalence of “probable iNPH” showed clear age dependency, ranging from 3.3/100 000 in the age group 50–59 years, 49.3/100 000 in the age group 60–69 years, 181.7/100 000 in the age group 70–79 years, to 93.3/100000 in the age group ≥80 years.

Three studies from Japan assess the prevalence of iNPH in a general population. Hiraoka et al. analyzed retrospectively data from a community-based study of 170 randomly selected residents aged 65 years or older, finding that 2.9% showed clinical and radiological features consistent with iNPH [16]. Tanaka et al. analyzed data from 1645 people aged 65 years and older in Tajiri, Japan [17], 497 of whom underwent MRI. Using radiological and clinical criteria, the authors identified 7 individuals with “possible iNPH,” yielding a prevalence of “possible iNPH” of 1.4% in this age group. No participants had the complete clinical triad. In an interesting prospective study, aiming to identify individuals having a preclinical stage of iNPH, 1142 residents aged 61 years and 70–72 years in two Japanese communities were invited to have brain MRI [18]; 790 (69.2%) participated. Four participants (0.51%) met the clinical and radiological criteria of “possible iNPH.” In addition, two out of eight participants with asymptomatic ventriculomegaly later developed dementia and/or gait disturbance during a follow-up of 4–8 years.

In a retrospective study, reviewing the Mayo Clinic’s medical records and available individual medical records from 1995 to 2003, Klassen and Ahlskog found 41 patients who had undergone an invasive diagnostic procedure for evaluation of suspected NPH [19]. This corresponds to an incidence of clinically suspected NPH of 3.74/100 000/year for the whole population, and 15.2/100 000/year for the population older than 50 years. Thirteen of these patients received shunts, yielding an incidence of NPH of 1.19/100 000/year when using “shunt insertion” as the final diagnostic criterion. The incidence of sustained definite improvements 3 years after shunting was 0.36/100 000/year. The authors did not differentiate between iNPH and secondary NPH (sNPH).

A study from Bangkok retrospectively reviewed the medical records of 293 patients with dementia or mild cognitive impairment assessed in an outpatient geriatric clinic. NPH was diagnosed in two patients (0.9%) [20]. The authors did not state the diagnostic criteria used for the diagnosis, nor did they differ between iNPH and sNPH.

Taken together, no firm conclusion can be drawn as to the true epidemiology of iNPH. The definition of the condition varies widely in the studies, ranging from no clearly specified definition [10,11], to strict definitions based on a good response from shunt insertion [8,12]. Many studies do not distinguish between iNPH and sNPH [8–10,12,19]. Also, several studies have taken into account only patients who have been admitted or evaluated for NPH [11,19], something that most certainly will underestimate the occurrence in the community-dwelling population. This is even more so in the studies that only take into account patients with a good response to shunt insertion [8,12]. On the other hand, the studies that estimate the occurrence of clinically and radiologically defined iNPH at a population or care facility level [7,9,13,14,16,17] will most certainly overestimate the occurrence of iNPH as long as a good response from shunt insertion is defined as a diagnostic criterion. This is because not all patients with clinically and radiologically defined iNPH will have a good response to shunt insertion. It can, however, be argued that a classification based only upon shunt response is a tautology, inherently excluding patients who have concomitant diseases that hinder a good shunt response, as well as patients in whom the pathological process has come too far to be reversed.

In line with all these difficulties in ascertaining the true epidemiology of iNPH, the estimates still differ widely in recent literature, leaving room for personal standpoints as to which epidemiological figures to emphasize from the international literature [19,21].

Secondary normal pressure hydrocephalus

The exact distribution of patients between the two conditions iNPH and sNPH is not fully known. In the previously cited multicenter study from Amsterdam constituting a total of 152 patients, 127 (84%) were classified as having iNPH, and 25 (16%) as having sNPH [8]. In a later Dutch multicenter study with 101 patients, 89% were classified as having iNPH, and 12% as having sNPH [22]. Nevertheless, some authors claim that the distribution between iNPH and sNPH is approximately 50/50 [23–25].

In vestibular schwannoma, the incidence of preoperative hydrocephalus has been reported to be as high as 18% (28/157) [26]. The hydrocephalus in this study was radiologically defined as mild in 11/28 (39%), moderate in 15/28 (54%), and severe in 2/28 (7%). Communicating hydrocephalus was present in 17/28 (61%) and noncommunicating in 11/28 (39%).

As to secondary hydrocephalus related to cerebral aneurysms, a nationwide American study conducted from 2002 to 2007 found hydrocephalus resulting in shunt placement in 9.3% of patients who underwent clipping after ruptured aneurysms, as compared to 10.5% in patients who underwent coiling [27]. The difference was not statistically significant. The frequency of hydrocephalus resulting in shunt placement after clipping or coiling in non-ruptured aneurysms was very much lower, 0.4% and 0.5%, respectively.

The frequency of surgery for iNPH and sNPH

In a Swedish study, the total incidence of surgery for adult hydrocephalus varied from 2.3 to 3.6/100 000/year between regions in Sweden during a 3-year period [28]. iNPH was listed as cause for operation in 243 out of a total of 891 operations, which equals an incidence of operation for iNPH of 0.91/100 000/year.

This figure is in line with the findings from a retrospective study conducted by the authors of this chapter and colleagues [29]. In this study, all operations for iNPH in Norway from 2002 to 2006 were registered, excluding patients who only underwent shunt revisions. A total of 252 patients were operated, giving a total incidence of 1.09/100 000/year. There was no clear trend as to whether the number of operations increased or decreased during the five-year period. The 5-year incidence was highest in the age group 70–79 years, reaching 52.1/100 000.

Finally, the previously cited work by Klassen and Ahlskog found an incidence of shunt insertion for NPH of 1.91/100 000/year, probably including both iNPH and sNPH [19].

These three studies all demonstrate a frequency of surgery for NPH that is far lower than the occurrence of iNPH found in most of the community-based studies with clinical and radiological criteria for iNPH. This discrepancy might have several explanations. iNPH is generally less known than many other diseases of the elderly, among both the public and general practitioners as well as other healthcare workers. The results from the information campaign we conducted in Vestfold, Norway, supports this, as it gave a total of 86 referrals and 20 shunted patients in one year, as opposed to a total of 18 shunted patients from the same region during the previous 10 years. This probably reflects a previous lack of awareness of the existence of the syndrome among healthcare workers. Adding to this is the fact that the symptoms of iNPH (gait problems, urinary incontinence, and dementia) are abundant in the elderly. Hence, even when they are aware of the syndrome, it is not necessarily evident for neither patients nor healthcare workers when and whom to refer. The lack of universally accepted guidelines for the diagnosis of iNPH, and the lack of powerful tests to predict shunt success, probably also contribute to the relative low rate of diagnosis, and consequently, of surgery. Lastly, iNPH mostly impinges on an age group where indeterminate, insidious symptoms often are considered an inevitable part of normal aging. A shuffling gait, or a gradual cognitive decline, might be inferred as a part of normal aging, and not as part of a treatable disease. The existence of iNPH thus reminds us that this is not always the case, and that such symptoms must be investigated in the individual patient, regardless of what one considers the “true” occurrence of hydrocephalic conditions.