Summary

7.1.1 Study setting and population

The data of this cross-sectional study was acquired from the semi-rural Inner-Savo district, which has a total population of 10,793 (Official Statistics of Finland 2017).

All people 65 years and older (N = 3,093) who lived at home and had diabetes were identified from primary care electronic patient records in primary health care.

According to the Social Insurance Institution of Finland, there were 540 persons aged 65 and older with reimbursed medication for diabetes in the Inner-Savo district in 2014 (KELA 2018).The diagnosis of diabetes was identified with the ICD-10 (World Health Organization 2018). The diagnoses were based on the criteria of the WHO and Finnish Current Care guideline for diabetes. Because only 12 people in the whole dataset had type 1 diabetes, people with type 1 and type 2 diabetes were combined. Two controls were matched by the same age and gender for each person with diabetes. Altogether 1,417 questionnaires were sent to 527 people with and 890 without diabetes. The participation rate was comprehensive (76.5%).

The health examination data of this study was based on a random sample drawn from the questionnaire study population including 259 people with and 259 without diabetes.

In general, compared with the study subjects without diabetes, people with diabetes had more chronic diseases, had depressive symptoms more often (especially women), smoked less often (especially men) and had less physical activity.

7.1.2 Prevalence of pain in older people

In older people both with and without diabetes, a high prevalence of frequent pain was discovered. Half of the participants had frequent pain. This can be roughly compared with the previous findings of prevalence of chronic pain varying widely between 30% to 74% (Fayaz et al. 2015, Rottenberg et al. 2015, Karttunen et al.

2015).

The present study found that instead of being associated with diabetes, pain is associated with the burden of diseases and depressive symptoms. This association

is in concordance with earlier studies (Baune et al. 2008, Bair et al. 2010, Caughey et al. 2010, Sudore et al. 2012, Jing et al. 2018).

However, women with diabetes had more frequent pain than women without diabetes. They had more osteoarthritis and pain in their knees and shoulders than women without diabetes. In earlier studies the difference in pain between women and men with diabetes has been associated with obesity and hyperlipidemia (McCarthy et al. 2008, Pico-Espinosa et al. 2017, Marttinen et al. 2019). In studies where woman and men were analyzed together, diabetes was related to chronic musculoskeletal-complaints (Hoff et al. 2008, Cole et al. 2009, Sowers et al. 2009, Zreik et al. 2016, Pico-Espinosa et al. 2017). In a Swedish study, diabetes was related to an increased risk for frequent back, neck and/or shoulder pain (Pico-Espinosa et al. 2017). Upper extremity pain and knee osteoarthritis are common in people with diabetes, and people with diabetes are more prone to have adhesive capsulitis than people without diabetes (Cole et al. 2009, Sowers et al. 2009, Zreik et al. 2016).

Because neuropathic pain could not be detected in this study, there is a possibility daily or continuous pain may be due to neuropathy. Based on the earlier studies, people with diabetes have more neuropathic pain (Davies et al.

2006, Abbott et al. 2011). Based on these studies, it could have been supposed that frequent pain in the present study could have been more prevalent. Thus, it can be assumed that in older people the etiology of neuropathic pain can include many other factors besides diabetes.

The response rate was considerably good, but it was better with diabetes than with controls (82% versus 74%). It is possible that those people who returned the questionnaire were different from those who returned. To assess this, a separate analysis was conducted. There were not differences between respondents and non-respondents regarding sex (p=0.16) or age (p=0.79). Therefore, it can be assumed that there was not a non-response bias that could have had a significant impact on the results.

7.1.3 Self-rated health

In the second part of this study it was found that frequent pain impairs SRH in older adults independently, which is in line in earlier studies. In another Finnish study, frequent chronic pain affected SRH independently and even more strongly than chronic diseases (cardiovascular diseases, diabetes, asthma) (Mäntyselkä et

independent effect on SHR and the number of chronic conditions made it worse (McDaid et al. 2013).

However, in the present study, no association was found between SRH and diabetes, contrary to the hypothesis that diabetes has an impact on SRH. Frequent pain may be important in allostatic load, which can be defined as the accumulation of impaired stress regulation and physiological adaptation (Sibille et al. 2017). The present findings may indicate that frequent pain increases allostatic load similarly in older people with and without diabetes.

7.1.4 The use of pain medication

It was supposed that people with diabetes may have more serious painful conditions than those without diabetes. Therefore, it was supposed that older people with diabetes may use more pain relieving drugs than people without diabetes. However, patterns of pain medication use seemed to be quite similar in older home-dwelling people with and without diabetes.

It was surprising that among the people having some pain, pain medication used as needed was 3.5 more common than regular use. This is a new finding and not in line with previous studies (Breivik et al 2006, Sarganas et al. 2015). This may represent the different therapeutic prescribing cultures. It may also indicate inadequate pain treatment and a need for proper treatment planning.

Previous studies have reported results which are generally concordant with the present study. However, probably due to different populations and study samples, the variation of the analgesic use has been quite large. The prevalence of analgesic users has varied between 22%-62% (Maxwell et al. 2008, Jyrkkä et al. 2009, Hamina et al. 2017). Previously, 15%-44% of the people with diabetes used pain medication (Wändell et al. 1996, Yang et al. 2019). Butchart et al. found that 74% of the people with and 78% without diabetes used analgesics (Butchart et al. 2009). This finding is quite close to the finding of the present study.

In the present study one out of seven people having pain medication used opioids. This is close to findings in earlier studies (Landi et al. 2001, Veal et al.

2015, Hamina et al. 2017). Despite having neuropathic pain and diabetes, few of them used neuropathic pain medication, which is concordant with the previous studies (Hartikainen et al. 2005, Pokela et al. 2010, Abdulla et al. 2013, Mills et al.

2016).

It could have also been justified to assume that due to their potential adverse effects, the use of NSAIDs in particular would have been less common in older people with diabetes. However, a significant difference was not found. Before

managing pain in older people, physician needs to know their patients´ history with morbidities and medication (Jones M et al. 2016). After assessing their pain individually, a physician can make a treatment plan for older people (Kaye et al.

2014). The aim of chronic pain management is to minimize the disadvantages of the pain and maximize the quality of life. Primary care has the ability to provide high-quality, easily available pain management (Breivik 2013, Davis et al. 2018).

7.1.5 Kidney function

This study found that there was no difference in kidney function between older people with and without diabetes and they had no increased risk for

nephrotoxicity related to the medicines used. Of the people with diabetes, 16%

had at least moderate decreased kidney function while the same number was 10%

in people without diabetes. This difference was not significant statistically. In additional analysis using the Cockcroft-Gault equation for eGFR taking into account weight, people with diabetes had a slightly but not statistically lower mean eGFR than people without diabetes. Nevertheless, these mean values are classified as normal renal function in both groups. In one study (Hobeika et al. 2015), older people with diabetes had the same baseline eGFR values as people without diabetes.

The prevalence of CKD in primary care older patients with diabetes type 2 is 20-50% wordwide (Thomas et al. 2006, Rodriguez-Poncelas et al. 2013, Lamine et al.

2016) and 18-35% in Finland in the same group (Metsärinne at al. 2015, Hagnäs et al. 2020). When screening only for GFR, the number of older people with diabetes having CKD was quite the same as ours, varying between 18-24% (Thomas et al.

2006, Lamine et al. 2016, Hagnäs et al.2020). When screening for albuminuria in addition to GFR, the numbers having CKD were higher 30-50% (Thomas et al. 2006, Rodriguez-Poncelas et al. 2013, Metsärinne at al 2015, Lamine et al. 2016, Hagnäs et al. 2020). According to diabetes classification, older people with mild age-related diabetes (MARD) had fewer metabolic disorders and a lower risk of chronic kidney disease than other groups (Ahlquist et al 2018).

In this study, the average number of potentially nephrotoxic drugs for each participant was 1.06 (±0.88) in people with diabetes and 0.97(±1.05) in people without diabetes. In a population-based study the mean number of nephrotoxic drugs per participant was almost the same, 1.2±0.6 (Breton et al. 2011). In a Finnish study of older home care clients with a heavier disease burden and older age, the mean number of medicines with renal risks was 2.4 per patient (Auvinen