The individual's age, sex, disease course, i.e. relapsing-remitting (RR), secondary progressive (SP), primary progressive (PP) and function, measured with the EDSS , together with date of onset of MS, were obtained from the SwMS register. The diagnostic criteria used in the SwMS register are both Poser criteria  (older cases) and McDonald criteria  (recent time). Secondary progressive course of MS (SP) is a clinical neurological status whose onset can be hard to pinpoint. In clinical trials and other studies, this status is often not defined in detail. In the SwMS register the following definition is chosen: A phase of the disease where one can secure a slowly increasing pyramidal para- or even tetraparesis, sometimes also a bladder syndrome. In some cases can it be a cerebellar syndrome. A crucial criterion is that you observe increasing pyramidal (or cerebellar) symptoms without any signs of bouts .
The patients' EDSS and disease-course are continuously upgraded at physicians call at the neurological policlinics. For the majority of the participants in this study, both the EDSS and disease-course had been determined within a year before the data collection started. Common background factors, such as civil status, family, level of education and ongoing medical treatment, were requested in the questionnaire. Occurrence of heat sensitivity was asked about in a single question, "Are you sensitive to heat?" (Yes/No), with follow-up questions concerning how this was experienced when in sunshine, in a warm room, or taking a hot bath or shower, and what room-temperature the patient preferred. Occurrence and perceived severity of fatigue were requested as well. Furthermore, the following instruments were included: the Fatigue Severity Scale (FSS) [7, 11]; the MS-related symptom checklist  and the Perceived Deficit Questionnaire (PDQ) .
In this study, and based on earlier studies [12, 14, 25], the EDSS was grouped according to normal neurological condition (EDSS = 0), mild disability (1.0≤EDSS≤3.5), moderate disability (4.0≤EDSS≤5.5) and severe disability (6.0≤EDSS≤6.5).
The FSS [7, 11] comprises nine items covering perceived severity of fatigue, and each item is graded from least fatigue (1) to severe fatigue (7). As in earlier studies [e.g. [9, 14], in this study an FSS mean score ≤4 was regarded as indicating no fatigue, >4 but <5 borderline fatigue, and ≥5 severe fatigue. The FSS has been used in earlier studies in Sweden  and is also used clinically. In this study, concurrent validity was assessed through correlations between the FSS with questions about the impact of fatigue on daily life as well as the occurrence of fatigue, rated on the MS-related symptom checklist , and resulted in r = 0.79, (P < 0.01) and r = 0.77 (P < 0.01), respectively. Reliability was assessed using Cronbach's alpha, and the alpha coefficient was 0.93.
The MS-related symptom checklist  was used to report 25 common MS symptoms such as fatigue, weakness in arms and legs, balance problems, pain, numbness, blurred vision, depression and difficulties with urination such as frequency and urgency. The occurrence of each symptom was graded in six steps from never (1) to always (6), which were then dichotomized, into never/sometimes (1-3) to 1 and usually/always (4-6) to 2. The checklist has been found to be closely related to the EDSS , and has been used in an earlier study in Sweden .
Cognitive dysfunction was assessed using the PDQ , measuring perceived problems with memory, attention and concentration. The original English version of the PDQ was translated into Swedish and then back-translated in accordance with Streiner and Norman , with good agreement. Twenty items are graded from never (0) to always (4). Summated scores vary from 0 to 80, where higher scores indicate greater cognitive problems. In this study, concurrent validity, assessed through correlations between the PDQ sum score and the MS-related symptoms forgetfulness and concentration difficulties, was r = 0.75 (P < 0.010) and r = 0.73 (P < 0.010), respectively. Reliability was tested using Cronbach's alpha (α = 0.95) and the split-half technique (r = 0.87).
The data have been treated and analysed in relation to perceived heat sensitivity. The descriptive statistics used are in congruence with measurement scale level, except for the FSS which is treated on interval level as in earlier studies [e.g. [9, 11, 14]. Differences between groups have been tested using the chi-square test on data on nominal level, the Mann Whitney U-test on data on ordinal level, and Student's t-test on data on interval level. To test associations between variables, Pearson's and Spearman's correlations were calculated. In the logistic regression analysis (enter), MS symptoms that were significantly more frequent among heat-sensitive participants were dichotomized and entered as dependent variables, and EDSS (0-6.5), disease course (relapsing-remitting/progressive forms), heat sensitivity (yes/no), time since onset, and age and sex (female/male) were independent variables. In the linear regression analysis (enter) mean FSS and summarized PDQ were entered as dependent variables and EDSS (0-6.5), disease course (relapsing/progressive forms), time since onset and heat sensitivity age and sex (female/male) were independent variables. In this study, and due to the number of statistical analyses, a P-value < 0.010 has been considered a significant value.