Risk factors of osteoporosis among adults in Ethiopia , the case of Tigrai region : A case control study

1 Department of Nursing, College of Health Sciences, Mekelle University, Ethiopia. 2 School of Public Health, College of Health Sciences, Addis Ababa University, Ethiopia. 3 Department of Economics, College of Business and Economics, Mekelle University, Ethiopia. 4 Department of Medicine, College of Health Sciences, Mekelle University, Ethiopia. 5 Department of medicine, College of Health Sciences, Mekelle University, Ethiopia.


INTRODUCTION
The patient profile in health institutions all over the developing world is changing.Non-communicable diseases (NCDs) have already established themselves as the predominant cause of disease and death in many middle-income countries including Ethiopia (WHO 2010).
Osteoporosis is one of the non-communicable diseases defined as a skeletal disorder characterized by low bone density and micro-architectural deterioration of bony tissue *Corresponding author.E-mail: alemayehub35@gmail.com.(Conference, 1991).It is one of the most common public health problems in adults and older people worldwide (Ross 1996).It is a common silent disease affecting both women and men over the age of 50 years.A significant number of clinical symptoms associated with osteoporosis become evident only after the occurrence of hip, vertebral, or wrist fractures.These fractures lead to many problems such as mortality, morbidity, and economic problems to individuals and the society (Bonjour 2001).
Previous studies have suggested that osteoporosis is a condition that can be prevented and treated if diagnosed early and accurately (Devine 1997).Unfortunately, it is often un-diagnosed until a fracture occurs; making the number of people who are screened for the disease very low.The correction of lifestyle factors including cigarette smoking and excessive intake of alcohol in preventing osteoporosis is a significant strategy that may be less costly when compared with medical and pharmacological interventions (Sharkey et al., 2000).Low bone mineral density in the adults can result from either low peak bone mass or accelerated bone loss, or a combination of the two.Nutritional factors play a role in both the attainment of peak bone mass and in the rate of age-related bone loss.Adequate calcium intake particularly milk and regular weight-bearing exercises are recommended for maintaining bone health and preventing excessive bone mass loss with advancing age (Eastell and Lambert 2002).
The exact prevalence of osteoporosis in developing countries is not well documented.However, according to the extrapolation of the prevalence of osteoporosis from developed countries and regions it is estimated that there are over 7.3 million cases of osteoporosis in Ethiopia (Census 2004).Despite this fact there has been a great interest in studying osteoporosis in developing countries.Currently, there has been no published study that examined the risk factors of osteoporosis in Ethiopian men and women.Hence, we undertook this study to determine the potential risk factors of osteoporosis in the Ethiopian context.

Study design
This is a hospital based, case control study with descriptive and analytic elements.

Study setting
The study was conducted in Tigrai region, Ethiopia, with urban inhabitants making 19.5% of the population.The region is predominantly Tigrai-Tigrinya people at 96.6% of the population, and 95.6% of the population are followers of the Ethiopian Orthodox.Public zonal hospitals are 12 in number and many other health centers are providing health services to the community.Currently, there is one specialized training hospital which provides advanced services to the community where the study is conducted and the digital x-ray machine is planted (CSA 2007).The study was conducted from May, 2012 up to October, 2012 on apparently healthy men and women aged 40 and above (the recommended age strata) who visited the regional referral and training hospital for any type of care where the Digital X-ray machine is situated were the source population.

Sample size
The sample size was calculated using two proportion formula for unmatched case control design.Assuming the proportion of cases with exposure rate of 27.51% (main exposure variable age people aged 50 and above) from studies conducted elsewhere (Sharma et al., 2006) and the odds of exposure (OR = 2.15) higher than controls, with 95% confidence interval and 80% power with case to control ratio of 1 to 2, a total of 133 cases and 266 controls were required.

Sampling technique
First, all relatively healthy adult men and women aged 40 and above who came to the referral hospital outpatient department during the data collection period were purposely approached to participate in the study.Second, radiographic examination of the wrist was conducted consecutively on those who volunteered to participate until the required sample size was obtained, for cases and controls, respectively.Pregnant women, chronically debilitated patients (with any known or unknown diseases), those with known diseases that affect bone metabolism (diabetes, hypertension, cancer, kidney diseases, etc.), and severely deformed patients (Kyphosis, anomalies of the anterior arm, etc.) were excluded from the study.

Ascertainment of cases and controls
Cases were osteoporotic patients that were screened by two senior radiologists after undergoing a digital X-ray scanning onto their nondominant wrist, while the controls were all participants without osteoporosis that were identified as osteoporosis free after undergoing the same procedure.Each X-ray result was visually inspected from the computer and read by the two radiologists independently; cases and controls were assigned when both agree.The X-ray results were discarded when they were considered to have inappropriate or poor quality by the two radiographers.X-ray results on which the radiologists did not agree were excluded from the analysis in this study.The percentage of agreement was 97.6%.

Data collection method
Data were collected using a pre-tested structured intervieweradministered World Health Organization (WHO) steps questionnaire for the identification of chronic disease risk factors (Strong and Bonita, 2002).Information on selected socio-demographic characteristics and lifestyle behaviors including physical activity, and anthropometric measurements of weight and height was obtained.Weight was measured with participants standing without shoes and wearing light clothing.Height was measured with participants standing upright and the head in the Frankfurt plane position.Height was recorded to the nearest 0.5 cm, and weight was recorded to the nearest 100 g.Body-mass index (BMI) was calculated as weight in kilograms over height in meters squared [weight (kg) / (height (m)) 2 ].Each participant was asked to provide information on current smoking habits and alcohol intake in average numbers of drinks per day at present.Coffee, milk intake, fruit and vegetable consumption were obtained.Medical history including personal and family history of fracture and previous use of steroids (predinsolone) were also acquired.
In this study, excess alcohol consumption is defined as an intake of any liquor (one bottle of beer, a glass of wine, spirits, fermented cider, Tella or Tej (Local beer)) more than 2 drinks a day.Consumption of coffee greater than 2 cups (70 ml) a day was also considered as excess.Furthermore, drinking less than five glasses of milk per week ( Heidi et al., 2003) and consuming fruits (orange, banana, etc.) and vegetables (Cabbage, spinach etc.) less than 5 times a week was considered as not adequate.In addition, when the work of the participant involves vigorous-intensity activity that causes large increases in breathing or heart rate like (carrying or lifting heavy loads, digging or construction work etc.) for at least 10 min continuously was considered as sufficient (WHO, 2011).The cutoff points are set after calculating the mean score of the continuous variables.

Data analysis
Data were entered and analyzed using statistical package for social sciences (SPSS) statistical software Version 19.Descriptive summary measures were used to depict levels of exposure in cases and controls.The associations between the exposure and outcome variables were determined using χ 2 statistics and backward stepwise binary logistic regression analysis to control confounders.Independent variables which have a P < 0.3 to the dependent variable in bivariate regression models were exported to a multiple regression model for sex and age adjustment.The major assumptions of logistic regression analysis (absence of multicollinearity and interaction among independent variables) were checked to be satisfied.The fitness of logistic regression models was assessed using the Hosmer-Lemeshow statistic and pseudo R squared values (Cox & Snell and Nagelkerke R squares).Crude odds ratios (COR) and adjusted odds ratios (AOR) along with their 95% confidence intervals (CI) were used to measure the degree of association and to test its statistical significance.

Ethical considerations
The study was conducted in line with national and international ethical recommendations to conduct biomedical research involving human subjects.Ethical clearance was secured from institutional review board of College of Health Sciences Addis Ababa University and an approval and letter of collaboration were solicited from Tigrai Regional Health Bureau.Informed written consent was taken from the study subjects after the objective of the study were briefed.Radiation safety procedures were in place.Osteoporosis related education was given to all study subjects.

Background information of the study subjects
A total of 396 participants-130 osteoporosis cases and 266 healthy controls participated in the study.Among which majority 83 (63.8%) of the cases and 135 (50.8%) of the controls were females, 75 (57.7%)and 134 (50.4%) of the controls were rural dwellers.Sixty seven (51.5%) of the cases and majority (89.5%) of controls were in the age group of 40 to 50 years.Furthermore, 23 (17.7%) of the cases and 39 (14.7%) of the controls had a family history of fracture after a minor bump or fall.There was a significant difference in age and gender of the participants with the p-value of 0.001 and 0.01, respectively (Table 1).

Lifestyle characteristics of the study subjects
Among the lifestyle variables the number of current smokers was few in both the cases and the controls.Alcohol intake ≤ 2 drinks per day was observed in 87 (66.9%) of the cases and in the majority (72.2%) of the controls.In addition, this study revealed that 101 (77.7%) of the cases and 204 (76.7%) controls consume coffee ≤ 2 cups a day.It was also observed that 85 (65.4%) of the cases and 90 (33.8%) of the control drink ≤ 4 cups of milk per week.Furthermore, 86 (66.2%) and 29 (22.3%) of the cases and 194 (72.9%) and 75 (28.2%) of the controls consume fruits and vegetables ≤ 4 times a week, respectively.The mean (SD) BMI of the cases and controls was almost similar.In addition, a significant difference was found in the weight of the study participants with a pvalue of 0.004 (Table 2).

Bivariate and multivariate analysis of the study subjects
The multivariate analysis adjusted for age and sex showed that rural residents were 1.93 times more likely to develop osteoporosis than the urban inhabitants with an adjusted odds ratio (AOR) 1.93 (95% CI, 1.11, 3.36).The strongest association was also found when the work of the respondent involves decreased physical activity with AOR 3.53 (95% CI, 1.98, 6.30).Furthermore, milk consumption greater than four times a week and smoking showed a significant association with the AOR 0.33 (95% CI, 0.19, 0.58), and AOR 0.17 (95% CI, 0.05, 0.58), respectively.Whereas, alcohol intake, coffee intake, fruit and vegetable consumption, steroid intake, BMI, occupation, education, monthly income, personal and family history of fractures were not associated with osteoporosis (Table 3).

DISCUSSION
We used unmatched case control study which is the first of its kind to identify risk factors of osteoporosis in the Ethiopian context.This study confirms some previously identified risk factors of osteoporosis such as consumption of milk, physical activity, current smoking status and residing in the rural.Normally we could expect that a  sedentary lifestyle which is more common in the urban population than is in the rural contributes to the development of osteoporosis, but, in this study rural residents were 1.93 times more likely to develop osteoporosis AOR 1.93 (95% CI: 1.11, 3.37), which might be related to the high prevalence of malnutrition in rural Tigrai.In this study, there was no significant difference between the social and demographic variables-education, monthly income, and family history of fracture and osteoporosis.This was discordant with findings in England, Turkey, Sweden, and Italy which revealed that osteoporosis was associated with poor education, family history of fracture, and low monthly income (Hui et al., 1988;Cankurtaran and Yavuz 2005;Olszynski et al., 2004;Biino et al., 2011).
Several studies in the USA reported positive associa-tions between earlier reported milk consumption or calcium intake and BMD, but several other studies did not find this association.The milk's main selling point is calcium, and milk-drinking is touted for building strong bones in children and preventing osteoporosis in older per sons (Heidi et al., 2003;Katherine 2003;Nguyen et al., 2000).Our study indicated that milk consumption up to four times a week has a protective effect of osteoporosis with the AOR 0.33 (95% CI: 0.19, 0.57) and this result is consistent with many other studies from India, UK, Iran, and USA (Heidi et al., 2003;Nguyen et al., 2000;Ruchira et al., 2010;Lunt et al., 2001).Evidence from the UK, USA, Australia, and India showed that exercise may help build and maintain bone density at any age.Studies have shown bone density increase by doing regular resistance exercises, such  as lifting weights two or three times a week.This type of weight-bearing exercise appears to stimulate bone formation and retain calcium in the bones that are bearing the load.The force of muscles pulling against bones stimulates this bone-building process.So any exercise that places force on a bone will strengthen that bone ( Jha et al., 2010;Todd Robinson 2003;Mozaffer et al., 2008;Forwood and Larsen 2000).The result of our study also supports these evidences; when the work of the participants does not include vigorous intensive exercise they are at risk of developing osteoporosis with an adjusted odds of 3.53 (95% CI: 1.98, 6.29) compared to their counterparts.Literatures from UK, Australia, USA and Europe also indicated that cigarette smoking is a risk factor for the development of osteoporosis; the reason is that nicotine and toxins in cigarettes affect bone health from many angles.Cigarette smoke generates huge amounts of free radicals molecules that attack and overwhelm the body's natural defenses.The result is a chain-reaction of damage throughout the body, including cells, organs, and hormones involved in keeping bones healthy (Law and Hackshaw 1997;Nguyen et al., 1994;Lunt et al., 2001;Daniel 1976;Williams et al., 2005.).Though the number of smokers is small, our study also revealed that nonsmokers are 0.17 times less likely to develop osteoporosis than the smokers with the AOR 0.17 (95% CI: 0.05, 0.58).
Moderate alcohol consumption is not harmful to bone health, but heavy drinking is a health risk for many reasons, including the effects on bones.Researches from Austria and Europe showed that chronic heavy alcohol use, especially during adolescence and young adult years by interfering calcium absorption, can dramatically affect bone health and increase the risk of osteoporosis later in life (Malik et al., 2009;De Vernejoul et al., 1983).However, this study observed alcohol intake had no significant association with the development of osteoporosis; this may be attributed to the fact that many people hesitate to tell the truth about the frequency and amount of alcohol intake due to high religiosity in the study area.Furthermore, it may be due to lack of statistical power of the study.
A study in USA implied that caffeine increases urinary calcium output and has been implicated as a risk factor for osteoporosis that leads to hip fracture (Douglas 1990), while some other researches in Sri Lanka, California, and Minnesota indicate that the effect of caffeine on bone mineral density is because of the interaction between cigarette and alcohol abuse (Glynn et al., 1995;Heaney 2002;Cooper et al., 1992).This study however revealed no significant association between caffeine intake and osteoporosis.Studies from the USA and Korea confirm that a high BMI (above 30 kg/m 2 ) has a protective effect for both men and women, in contrast low BMI less than 19 kg/m 2 can lead to osteoporosis (Wardlaw 1996;Barrera et al., 2004;Kenny et al., 2000;Shin et al., 2004).While our study failed to show the association between BMI and osteoporosis.A similar reason can be given that the difference in mean body weight and mean height of the cases and the controls was almost similar.
A study in the USA has linked higher intakes of fruits and vegetables with better bone health.Though it is not clear why fruits and vegetables promote healthy bones, many scientists believe that fruits and vegetables contain certain nutrients such as calcium, magnesium, potassium, vitamin C, vitamin K or a combination of these vitamins that are beneficial for bones.Some studies indicate that the higher fruit intake was found to be significantly associated with higher BMD in both sexes.High vegetable consumption, however, did not positively impact BMD Zalloua et al., 2007.Further UK and USA studies also indicate consumption of more fresh fruits and vegetables is unlikely to be detrimental to bone health and may be beneficial (Prentice 2004;Vatanparast et al., 2005).However, the current study did not show an association between fruit and vegetable consumption and osteoporosis, under reporting of dietary intake by the participants might explain why this difference was not apparent in the study.
Corticosteroids have several adverse effects on bone metabolism.Direct inhibition of osteoblast function, direct enhancement of bone resorption, inhibition of gastrointestinal calcium absorption, increase in urinary calcium loss, and inhibition of gonadal hormones mainly affect the trabecular bone.In our study, corticosteroid intake did not show a significant association, this could be because of the problem of reporting the exact type of medication (Walsh et al., 2002;Sinigaglia et al., 2000;IP et al., 1994) Studies from UK and USA confirmed that family and personal history of osteoporosis or osteoporotic fracture is a risk factor for osteoporosis (Kanis et al., 2004;Ralston and de Crombrugghe, 2006), while our study found a lack of association between family and personal history of fracture among the cases and controls.The reason for the lack of association is not clear.

LIMITATIONS OF THE STUDY
Though the study is the first of its kind in the region the results have to be interpreted within the context of potential limitations.First, the population is composed of volunteers not randomly selected who do not ensure representation of general population; therefore, this may introduce selection bias.Second, because we excluded individuals with diseases deemed to interfere with bone metabolism, the risk factors of osteoporosis reported here could be an underestimate of the risk factors in the general population.Third, we have used mean scores for milk and fruit and vegetable consumption for comparison with other literatures; this should have to be interpreted cautiously.Fourth, we did not use the gold standard diagnostic DEXA to assign cases and controls that might lead to inflate the false negative rate.However, this research did not intend to establish the prevalence of osteoporosis but to identify associations with low bone mass (osteoporosis).On the other hand, our study is the first to report risk factors of osteoporosis in men and women in Ethiopia, while many of the published studies in the literature come from American, European and Asian countries.

Conclusion
Residing in the rural setting and smoking was positively associated with osteoporosis.In contrast, milk intake greater than four times a week, and when work involves vigorous exercise appeared to be associated with a reduced risk of osteoporosis.Therefore, the findings from the study suggest the need for changes of lifestyle that includes encouraging adults to stop smoking, engage in vigorous physical activities for active aging and adequate dietary intake including milk.Strategies to identify and manage osteoporosis in the primary health care setting to reach the rural people need to be established.Large community-based studies using the gold standard bone density scanner for a better knowledge of the risk factors of osteoporosis for early prevention and treatment is needed.

Table 1 .
Socio-demographic and fracture history related variables of the respondents by their osteoporosis status, Tigrai, Northern Ethiopia, 2012 (n=396).

Table 2 .
Lifestyle, anthropometric, nutrition and drug related variables of the respondents by their osteoporosis status, Tigrai, Northern Ethiopia, 2012 (n=396).
All variables adjusted for age and sex, COR= Crude odds ratio (Unadjusted odds ratio), AOR = Adjusted odds ratio, *Significant association.