in middle-income countries.

1470 www.thelancet.com Vol 380 October 27, 2012 Women who have undergone genital mutilation cannot be compared to individuals with intersex conditions (Creighton and colleagues’ references 2 and 3), in terms of either psychological experience or the type of lesion to be repaired. We clearly set out the limitations of the study in our paper. We are aware of Creighton and colleagues’ strong opposition to female genital cosmetic surgery. However, what we are doing can in no way be described as “cosmetic”. Instead, what we are trying to do is to restore the dignity and wellbeing of women who have experienced violence—and who very much welcome our initiative.

Women who have undergone genital mutilation cannot be compared to individuals with intersex conditions (Creighton and colleagues' references 2 and 3), in terms of either psychological experience or the type of lesion to be repaired. We clearly set out the limitations of the study in our paper.
We are aware of Creighton and colleagues' strong opposition to female genital cosmetic surgery. However, what we are doing can in no way be described as "cosmetic". Instead, what we are trying to do is to restore the dignity and wellbeing of women who have experienced violence-and who very much welcome our initiative.
We declare that we have no confl icts of interest. "might be much the same as that typically recorded in countries with high incomes". 1 A possible explanation for this apparently negative result might be found in the analysis of dementia incidence by age. In a retrospective study on dementia prevalence in Latin America, 3 we found a similar total dementia prevalence to that of studies from developed countries, but the prevalence was higher in relatively younger individuals and lower in older individuals in Latin America than in developed countries. According to the cognitive reserve hypothesis, higher cognitive reserve delays the clinical manifestations of the diseases that cause dementia. 4,5 Our interpretation was that low cognitive reserve caused earlier emergence of clinical signs of dementia in Latin America, whereas high mortality decreased the prevalence in older individuals. 3 It is not clear whether Prince and colleagues assessed the possibility of diff erent incidence by age in their sample when compared with studies from high-income countries. Information on this point might be important for public health initiatives in low-income and middle-income countries.

P Foldès, *B Cuzin, A Andro
I declare that I have no confl icts of interest. Martin Prince and colleagues 1 classifi ed patients with dementia according to rural and urban residence without mentioning diff erences in exposure to air pollution. Studies have indicated associations between air pollution and neurodegenerative diseases. Residents of cities with severe air pollution had signifi cantly greater accumulation of β amyloid protein in the brain than did residents of less polluted cities. 2 Research implies that elevated levels of air pollutants could have a role in neuroinfl ammation and the pathogenesis of neurodegenerative disorders. Consequently, air pollution should be considered as a confounding factor in the study by Prince and colleagues.

Ricardo
Research has also revealed links between environmental inhalation exposures and olfactory function in human beings. Signifi cant olfactory impairment associated with exposure to air pollution has been reported. 3 Olfactory loss is an early fi nding in neurodegenerative diseases, pre ceding cognitive and motor symptoms by several years. 4 Also, olfactory dysfunction was found to be correlated with β amyloid burden in a mouse model of Alzheimer's disease. 5 Linking these study results together suggests the question: could a causal relation exist between poor air quality and neurodegenerative diseases, with impaired olfaction being an early marker?
Thus, the use of periodic olfactory testing as an early screening tool for possible neurotoxicity after substantial air pollution exposures could be clinically useful. Confounding factors, such as ageing or vascular disorders that aff ect olfaction, are prevalent in older people. Therefore, young people are recommended for future research.
We declare that we have no confl icts of interest.

rfphalen@uci.edu
Department of Medicine, University of California, Irvine, CA 92617, USA

Dementia incidence in middle-income countries
One of the main hypotheses in Martin Prince and colleagues' study on dementia incidence in middleincome countries (July 7, p 50) 1 is that if cognitive reserve is less stimulated in these countries, dementia incidence should be higher than in high-income countries. Considering that education is one of the main constituents of cognitive reserve, 2 and that in the total baseline sample of the study 41% did not complete primary education, one indeed should expect a higher dementia incidence in this sample. However, the results of the study suggest that dementia incidence countries, with the highest level of treatment at 14%. These results point not only to treatment gaps, but also care insuffi ciencies. Not addressing this risk factor is likely to result in high disease prevalence and costs from chronic treatment of cerebrovascular and cardiovascular disease. Yet treatment of hypertension is highly cost eff ective. 5 SAGE provides an evidence base to track hypertension over time, as well as related health expenditures and covariates that aff ect prevalence, incidence, and treatment.
We declare that we have no confl icts of interest.

Hypertension in developing countries
The Series paper by Mohsen Ibrahim and Albertino Damasceno (Aug 11, p 611) 1 points to a need for action and research on hypertension in lowincome and middle-income countries.
Relatively little is known about prevalence in older adults. One source of health data for six countries, WHO's Study on global AGEing and adult health (SAGE), found hypertension to be common in older adults, with a higher prevalence in women and urban residents. 2,3 For those with multiple chronic disorders, a high proportion include hypertension, 3 adding to disease burden and treatment complexities. The prevalence and awareness of hypertension in urban and rural dwellers in SAGE wave 1 (China, Ghana, India, Mexico, Russia, and South Africa) 4 show marked diff erences in those with diagnosed hypertension under adequate control by age and urban or rural residence (fi gure). Individuals not diagnosed but with high blood pressure on measurement (higher in rural settings) are as much of a concern as those who know they have hypertension and are still hypertensive on measurement (much higher in urban settings). Awareness was low in four of the six Mohsen Ibrahim and Albertino Damasceno 1 describe the increasing preva lence of hypertension in developing countries and propose measures for blood pressure control. Meanwhile, rates of cardiovascular morbidity and mortality are also growing in these regions. 2,3 It is important to understand the relation between hypertension and cardiovascular disease in developing countries. In a traditional rural population in northern Ghana, we have studied hypertension, peripheral arterial dis ease, and coronary arterial disease, by use of ankle-arm index and electrocardiography, in almost